Lluis Gene / AFP / Getty Images

A spokesperson for Facebook, accidentally responding to a BuzzFeed News reporter via email, called the notion of a Muslim registry a "straw man." Seemingly thinking he was addressing a colleague, he suggested that the best course of action was to not respond to the reporter's inquiry.

Earlier today BuzzFeed News emailed Facebook to ask whether the social networking giant would make a commitment to limit data collection that could be used for ethnic or religious targeting, including a pledge not to build a registry of Muslims, if asked to do so by the government. A Facebook public relations representative intended to forward our request, along with a message about how to respond, within Facebook, but accidentally sent the email to BuzzFeed News instead and in doing so, provided inadvertent insight into how the company plays the optics game.

BuzzFeed News

Facebook had not responded to previous requests for clarification on whether or not it would participate in building a Muslim registry. BuzzFeed News asked again today in light of the fact that Facebook COO Sheryl Sandberg will attend a tech summit at Trump Tower tomorrow, and because a group of 60 or so engineers and employees for major tech companies just signed a pledge not to comply with practices that could be used to target people or build databases based on their religious beliefs.

In the email that was accidentally sent to BuzzFeed News, the PR rep calls this "attacking a straw man."

Happy to talk to her off record about why this is attacking a straw man. Also I heard back from her that she may or may not write an additional piece depending on what response she gets from companies. So sounds like not making any stmt on record is the way to go.

The representative subsequently called, and asked that the email be considered off the record. This preference for off the record spin over on the record comments is fairly typical of large tech companies. Facebook ultimately declined to comment.

However, the possibility that President-elect Donald Trump will try to create a Muslim registry is not a straw man. Trump has repeatedly been given a chance to clarify whether he supports the idea. He has been asked about it by everyone from Fox News to George Stephanopoulos. On more than one occasion he has chosen not to rule out the possibility of a Muslim registry. The question of such a registry arose based on Trump's comments about Muslims and his intention to build a database of Syrian refugees.

Since the election, Kansas Secretary of State Kris Kobach, an advisor to Trump's transition team, told Reuters that Trump's immigration policy group is considering reinstating the National Security Entry-Exit Registration System (NSEERS), a suspended federal program used to keep a database of immigrants from countries that have a Muslim majority from 2001 to 2012.

Facebook, of course, already asks for and retains sensitive information about the race, religion, and location of its users and allows advertisers to target narrow segments of people based on that personal information. Government officials here and abroad already use the social network to track activists and dissidents.

View Video ›

Facebook: video.php

On the same day Uber launched a pilot program to put riders in its self-driving cars in San Francisco, state regulators have asked the company to hit the brakes.

The California Department of Motor Vehicles sent a letter to Uber Wednesday asking the company to cease its program until it obtains a permit from the agency. The California DMV requires companies looking to test autonomous vehicles to obtain a permit and submit accident reports to the agency. But because its cars will have human drivers, Uber says it doesn't need a permit.

On Wednesday, one of Uber's self-driving cars was caught on camera running a red light in San Francisco.

"It is illegal for the company to operate its vehicles on self-driving roads until it receives an autonomous vehicle testing permit," the letter, obtained by BuzzFeed News, reads. It adds that Uber "must cease" testing its self-driving cars on public streets until it receives a permit from the DMV.

"If Uber does not confirm immediately that it will stop its launch and seek a testing permit, DMV will initiate legal action," the letter reads.

Uber did not immediately return a request for comment on the DMV's letter.

Uber began putting riders in autonomous cars in Pittsburgh in September as part of a pilot program. Starting Wednesday, riders who hail an UberX to travel within San Francisco's limits might be picked up in one of the company’s handful of self-driving Volvos. Uber wouldn’t say exactly how many self-driving vehicles hit the road in San Francisco on Wednesday, but said it plans to have a total of 100 self-driving cars there and Pittsburgh by the end of 2016.

For now, the cars aren’t 100% autonomous: Each vehicle still has a human driver behind the wheel to begin each ride — by starting the ignition, shifting gears, and exiting a parking lot to reach the road, for example. A screen on the dashboard will tell the driver when the car’s computers determine it is safe to turn on autonomous mode. And in the front passenger seat, a human co-pilot with a laptop will monitor the car’s trajectory. The driver can take over at any time, and if the car determines a situation is too precarious, it may also beep and kick control back to the driver, whose hands are to remain hovered near the wheel. There’s a big red button near the gears, which a driver can push to turn off autonomous mode, or he or she can simply resume control of the vehicle by taking the wheel.

If you’re matched with a self-driving Uber, you’ll see this notification:

Uber

Riders have the option to turn down autonomous Uber rides by canceling their requests.

Unlike several of its competitors, including Google and Tesla, Uber is notably absent from a list of 20 companies that have obtained permits from California's Department of Motor Vehicles to test autonomous cars.

"First, we are not planning to operate any differently than in Pittsburgh, where our pilot has been running successfully for several months," Anthony Levandowski, head of Uber’s self-driving team, wrote in a blog post announcing the SF expansion. "Second, the rules apply to cars that can drive without someone controlling or monitoring them. For us, it’s still early days and our cars are not yet ready to drive without a person monitoring them."

California's rules say "a motor vehicle shall not be operated in autonomous mode on public roads" without a permit. And according to the rules, "an autonomous vehicle is operating or driving in autonomous mode when it is operated or driven with the autonomous technology engaged."

Uber is arguing that its human drivers still have active control, and that there is a difference between "self-driving" and the word "autonomous."

Driving in San Francisco will present a different set of challenges for autonomous cars than Pittsburgh: The California city has more traffic, more bikes and pedestrians, and narrower lanes. The self-driving cars will only accept rides from passengers whose routes are contained within the city of San Francisco’s perimeters – so no trips to Oakland or Palo Alto, for example.

“It will look for routes where we have excellent support for autonomy,” said Matt Sweeney, head of product at Uber. The cars will rely on Uber’s own mapping technology and will drive at a maximum speed of 30 mph, meaning they will avoid highways.

Sweeney wouldn’t say which cities Uber is planning to bring its autonomous vehicles to next, but its ambitions for self-driving vehicles are expansive.

“The promise of self-driving is core to our mission of reliable transportation, everywhere for everyone,” Uber's Levandowski wrote in a blog post about the program. “While it won’t happen overnight, self-driving will be an important part of the future.”

The company became the first to put passengers in self-driving cars through ride-hail in September in Pittsburgh, only 18 months are opening its Advanced Technologies Center, where many of its autonomous driving engineers work. Uber also purchased Otto, a self-driving truck startup, this summer. Otto completed its first self-driving truck delivery — a beer run for Anheuser-Busch — in October. And earlier this month, Uber announced it had purchased an artificial intelligence startup called Geometric Intelligence and said it is opening an AI lab to focus on machine-learning efforts such as self-driving.

Uber’s self-driving expansion to SF also makes its rivalry with Google, based in nearby Mountain View, more evident. Google announced on Tuesday that it’s spinning its self-driving car program off into an independent company called Waymo, which will live under the parent company Alphabet’s umbrella. Waymo is also planning a ride-sharing program with Fiat Chrysler, The Information reported Monday.

Brian Ach / Getty Images

Amazon has launched its video streaming service Prime Video in 200 countries, including India, the company's largest geographical market outside the United States.

The service is bundled with a Prime subscription, and in some countries where Prime isn't available, it will cost $5.99 on its own -- slightly cheaper than the $8.99 it costs as a standalone service in the United States.

This move establishes Amazon Prime Video as a global competitor to Netflix, which expanded simultaneously to 130 countries in January.

In price-conscious markets like India, a Prime subscription is ridiculously cheap compared to $99 a year in the US. Indians can get Prime at Rs. 999 a year ($15) , and it is available at half that price currently to encourage new sign-ups. The cheapest Netflix subscription by contrast costs Rs. 500 a month ($7.4) in India.

Expanding Prime Video globally is a strategic milestone for Amazon, but it is especially relevant in countries like India, which are important markets for most US tech companies thanks to rapid smartphone penetration and falling data prices.

At an event in Washington in June, Amazon CEO Jeff Bezos said that the company would invest up to $5 billion in India. Amazon's aggressive expansion in the country has worried local rivals like Flipkart, India's largest e-commerce startup, which recently appealed for protectionist policies from the Indian government to stave off cash-rich rivals like Amazon.

Amazon is counting on localisation to compete with Netflix. For instance, Netflix had no Indian content at all when it launched in country. However, Amazon Prime Video already has dozens of launch titles in Hindi, Bengali, Marathi, Tamil, and Telugu. Similarly, in the European markets, Prime Video content has French, Italian, Portuguese and Spanish subtitles.

Amazon has also partnered with several major film production studios in India to make dozens of Bollywood titles available on its service, and has included local language subtitles for English-language shows like Mad Men, Seinfeld, Person of Interest, and more. Also available are US-produced Amazon Originals series like Transparent and Bosch.

Jacoblund / Getty Images

More than a year and a half ago, Apple unveiled a new breed of iPhone apps that would let people participate in scientific studies anytime, anywhere — at least in theory. Now, a new study indicates that smartphones do have the potential to capture useful data about thousands of people’s health and exercise habits in their daily lives, not just during a trip to the doctor or a clinical trial center.

But it also shows that if iPhones are to become the next big tool in science, researchers will have to conquer a challenge familiar to every app developer: how to keep people from getting bored and quitting.

The study, published Wednesday in JAMA Cardiology, reveals the first results of an app that was part of Apple’s much-hyped foray into health, a field that has been relatively slow to digitize. “This is the blossoming of mobile device medical research,” said Eric Topol, a cardiologist and genetics researcher at the Scripps Research Institute, who was not involved with the study, in an interview with BuzzFeed News.

At a widely publicized and livestreamed event in March 2015, Apple introduced ResearchKit, a tool to build apps to study conditions like breast cancer, Parkinson’s disease, type 2 diabetes, and asthma — to name a few of the ResearchKit apps that launched that month.

MyHeart Counts

Via itunes.apple.com

One of the apps, MyHeart Counts, run by a team at Stanford University, aimed to measure people’s cardiovascular health. Between March and October 2015, nearly 50,000 people from all 50 states signed up for the weeklong study, and 40,000 of them submitted health data of some kind. These unusually high numbers suggest that smartphone-conducted studies can reach many more people than traditional studies ever have.

Before this, “there hasn’t been 40,000 patients capturing their data about their activity through a medical research app,” Topol said. “That is a standout.”

At the same time, relatively few people took the extra step of doing some of the tasks that the app asked them to do. Just shy of 5,000 people completed a six-minute walk test, a common proxy of heart health, with their phones in hand, according to the study. And while 40,000 people filled out some portion of the app’s health questionnaires, only about 1,300 provided all the information needed to calculate their personalized health risks.

That squares with a recent study that showed that for the first five ResearchKit apps that launched, including MyHeart Counts, the percentage of daily users quickly dropped to 25% or below within the first three months.

“This is a very significant problem with this new form of medical research,” Topol said. “You accrue lots of people, but to keep them engaged, long-term, is perhaps the greatest challenge.” Apps could keep people coming back, he suggested, by offering participants something valuable in return, like personalized insights about their health.

While Euan Ashley, the study’s senior author, acknowledges the steep drop-off rate, he sees those figures optimistically. Although only 10% or so of eligible people did the six-minute test, Ashley noted, this group was about 10 times bigger than any other study of people doing the same test. Clinical trials in general sometimes see 30% or more of patients drop out.

The app was also valuable, he said, because it monitored people’s physical movements in the background. It showed, for example, that people were active about 15% of the time their movements were recorded in a week (assuming that people carried their phones or wore a fitness-tracking device when on the go).

“Instead of asking ‘How many steps did you take last week?’, ‘How many flights did you climb?’, we can actually measure that. I think that’s a groundbreaking thing.”

“The scope of it and scale of it is new and represents a huge opportunity,” Ashley, an associate professor of cardiovascular medicine at Stanford University School of Medicine and one of the leaders of the MyHeart Counts study, told BuzzFeed News. “Instead of asking ‘How many steps did you take last week?’, ‘How many flights did you climb?’, we can actually measure that. I think that’s a groundbreaking thing.”

When MyHeart Counts and the other original ResearchKit apps launched, no one — including researchers — knew whether people would be receptive, Ashley said. Since then, several more ResearchKit apps have launched.

“Given that these launch studies were the first of a kind, the idea you could use these devices that fit in everyone’s pockets and pull them out multiple times a day for clinical research to help patients — that was unknown,” he said. “The number-one finding that we’re reporting is that this is possible.”

While all clinical trials struggle to enroll a diverse and representative group of patients, this study’s demographics also underscore some inherent challenges to smartphone studies. Obviously, only people with the financial means to own iPhones could participate. And while clinical researchers traditionally meet their subjects in person, that’s not possible when the study is done exclusively through an app. So it’s possible that a small number of people purposely faked their data, Ashley acknowledged, although he doubts that the group was big enough to skew the data.

Christine Lemke, president of Evidation Health, a startup that studies and validates health apps, called the study a promising look at the potential future of research.

“It’s feasible to deliver studies like this,” she said. “More of them, please.”

Toru Hanai / Reuters

President-elect Donald Trump named Elon Musk, the head of SpaceX and Tesla, and Uber CEO Travis Kalanick to his growing panel of industry chieftains Wednesday, a group convened to provide regular economic guidance to the next president. PepsiCo's CEO Indra Nooyi has been added as well.

Billed as the President’s Strategic and Policy Forum, the panel was announced earlier this month, comprised of 16 business leaders, including the heads of Walmart, Disney, and General Motors. Among the initial batch of CEOs, IBM was the sole technology company represented, making the addition of Musk and Kalanick, all the more important in Trump's efforts to channel and recruit the expertise of Silicon Valley.

“America has the most innovative and vibrant companies in the world, and the pioneering CEOs joining this Forum today are at the top of their fields,” Trump said, when the forum was first unveiled. “My Administration is going to work together with the private sector to improve the business climate and make it attractive for firms to create new jobs across the United States from Silicon Valley to the heartland.”

Tesla has not yet responded to a request for comment.

Kalanick told BuzzFeed News in a statement: "I look forward to engaging with our incoming president and this group on issues that affect our riders, drivers and the 450+ cities where we operate.”

Trump is slated to meet with Musk and a who's who of tech industry titans Wednesday afternoon, with job creation expected at the top of the agenda. Among the guests planning to attend: Tim Cook of Apple, Jeff Bezos of Amazon, Sheryl Sandberg of Facebook, and Satya Nadella of Microsoft.

Cook and Musk are also expected to meet with Trump privately. Apple has not responded to a request for comment.

Facebook’s “M” virtual assistant hasn’t been rolled out to all that many people, but its interactions with a limited user base have helped train Facebook’s artificial intelligence systems, and now the masses may benefit.

M suggestions in action

Today, Facebook is starting a small test in the US for “M suggestions,” a new feature that will suggest certain actions based on the context in conversations within the Messenger app.

The test feature will suggest things like sending a location after someone asks a question like “where are you?,” or it will offer a small selection of stickers you may want to send in response to a message. “M” will show up in these conversations with its own chat avatar.

“Think of this as a version of M that can actually help suggest the right capabilities at the right time,” Messenger head David Marcus told BuzzFeed News.

The virtual assistant version of M has been in tests for over a year and still has no set date for a bigger rollout, but Marcus said Facebook wants to bring it to the public. “M suggestions,” if rolled out more broadly, could help push the schedule up a bit, getting a completely automated version of M into the hands of Messenger's over 1 billion users, albeit in a much more lightweight capacity than the assistant version.

“Hopefully with this side of it, we’ll have a path to opening it up to everyone fairly quickly,” Marcus said.

There’s still a good deal of technology that goes into creating these basic suggestions. Facebook essentially needs to understand what you’re saying in messages to make sure its suggestions aren’t annoying, and its technology also adjusts its suggestions based on how you interact with M. “It seems completely pedestrian, but it’s actually very hard to do,” Marcus said, referencing Messenger’s sticker suggestions.

Other suggestions Messenger already makes, such as event reminders and ride hailing, will be rolled under the “M suggestions” umbrella as well.

Facebook M in action

Facebook's M suggestions can be seen as a defensive move in some ways. Google Assistant, a virtual assistant inside Google’s Allo messaging app, can also be called into conversations, and Allo suggests responses based on the conversation’s context with “Smart Replies.” Asked for his thoughts on Google Assistant, Marcus declined to address it directly.

The virtual assistant competition between Google and Facebook could get interesting very quickly, especially since assistants fit in both companies’ territory. A great assistant could be a competitive advantage for a messaging app, and if it's developed outside of Facebook, it could be a threat since the company owns both Messenger and WhatsApp, two massive messaging apps. On the other hand, virtual assistants can help people look for something online, the pillar Google’s business is built on. For now though, neither company appears to be under imminent threat from the others’ virtual assistant efforts.

M suggestions will roll out to a very small group, but Marcus said Facebook hopes to gradually expand it in the first half of next year.

Carl Court / Getty Images

The owner of AshleyMadison.com will pay a fraction of a $17.5 million settlement reached Wednesday over the hack of the cheating website in July 2015 that led to the release of millions of members' personal information.

Ruby, which owns Ashley Madison in addition to CougarLife and Established Men, will be required to make an immediate payment of just over $1.6 million divided among 13 states included in the suit and the Federal Trade Commission. But "the remainder of the $17.5 million payment is suspended based on ruby Corp.’s inability to pay," the New York Attorney General's office said in a statement.

The office told BuzzFeed News settlement figures are not determined by a company's ability to pay but on the violations they committed. “This settlement should send a clear message to all companies doing business online that reckless disregard for data security will not be tolerated,” said Attorney General Eric Schneiderman in a release.

The discussion of Ruby's ability to pay the $17.5 million settlement is not public, and the company's recent financials were not immediately available. In 2014 — before the site was hit by security concerns — the owner of Ashley Madison made $115 million in revenue. If the office discovers the company misrepresented their finances, Ruby would be required to pay the remainder of the $17.5 million.

FTC Chair Edith Ramirez told reporters in a press call on Wednesday that the goal of the settlement was to make the company "feel the pain."

"We don't want them to profit from unlawful conduct," she continued. "And at the same time we are not going to seek to put a company out of business."

The FTC's portion of the settlement will go to the US treasury, and not to consumers, because the $1.6 million figure would not cover compensation for those affected, added Ramirez.

View Video ›

Ashley Madison / Via Facebook: AshleyMadison

An investigation into the hack found AshleyMadison.com had weak security practices, misrepresented its security to users, and failed to delete user information after they paid to be erased from the site, said the New York Attorney General's office.

The site also created fake profiles to lure users who were using Ashley Madison’s free services into purchasing credits to communicate with other members, including “members” with other fake profiles, investigators found. Some of these fake profiles used information and photos from actual users who had been dormant for a year.

Ruby said in a statement that the settlement marks a "pivotal day" for members and the company. But it "neither admits nor denies the allegations made by the FTC and the State Attorneys-General."

"Today’s settlement closes an important chapter on the company’s past and reinforces our commitment to operating with integrity and to building a new future for our members, our team and our company,” said Ruby CEO Rob Segal, adding the company has been cooperating fully with the FTC for more than a year.

“Our team is making significant, long-term investments in our people, processes and systems to improve Ashley Madison for our members,” he said.

Ruby's president James Millership said the company has already implemented new business practices and improved its security over the last 18 months.

Leaked Emails Suggest Ashley Madison Founder Had Multiple Affairs

That Ashley Madison Hack Just Got Much, Much Bigger

Ashley Madison’s $19 ‘Full Delete’ Option Made The Company Millions

That feeling when.

Drew Angerer / Getty Images

Evan Vucci / AP

Shannon Stapleton / Reuters

Drew Angerer / Getty Images

View Entire List ›

Yahoo Mail logo is displayed on a smartphone's screen in front of a code in this illustration taken in October 6, 2016.

Dado Ruvic / Reuters

Yahoo believes, based on third party forensic evidence, that someone stole user data from one billion Yahoo accounts in 2013, according to a press release from the company. This theft, it says, is likely separate from the hack of 500 million accounts it disclosed in September.

The company received information about the breach from law enforcement on November 7, 2016, according to a spokesperson. But the company has not been able to identify the perpetrators of the billion-account hack, the press release said. Yahoo is requiring all potentially affected users to change their passwords and password recovery questions.

The data may have included “names, email addresses, telephone numbers, dates of birth, hashed passwords (using MD5) and, in some cases, encrypted or unencrypted security questions and answers.” Credit card info was not affected, according to Yahoo.

Flickr users should also change their passwords and otherwise ensure their accounts are secure.

In early August 2016, a person on the Dark Web going by the name "Peace" listed data from 200 million Yahoo accounts for sale. At the time, Yahoo said it was aware of the listing, but it did not issue a password reset.

A source close to the investigation told BuzzFeed News that Yahoo looked into Peace’s claim in July 2016 and found no direct evidence to substantiate it. According to the source, however, Yahoo then began another, wider investigation that led them to discover the breach of 500 million accounts.

In October 2016, Reuters reported that Yahoo had built custom software for US intelligence to spy on Yahoo Mail users, which also incurred the ire of cybersecurity experts and Yahoo users alike, though Yahoo denied the existence of the software.

This new hack, especially when considered with the 500 million account hack in September, may affect the proposed $4.8 billion sale of Yahoo’s core business to Verizon. Reuters earlier reported that Verizon pushed for a $1 billion discount on the deal.

Verizon reiterated its statement from the September hack: "As we've said all along, we will evaluate the situation as Yahoo continues its investigation. We will review the impact of this new development before reaching any final conclusions."

Arun Vishwanath, a professor of communications at the University of Buffalo who studies cybercrime, said that since many people use one password for multiple accounts, the Yahoo hackers are likely attempting to breach other accounts across the internet.

"The most likely thing hackers have done is harness the passwords to affect targeted attacks or for conducting widespread DDoS type attacks using the known password combinations, as we saw with the recent Dyn attack," Vishwanath said.

He recommends changing your passwords, enabling multi-factor authentication, and logging out after each email session.

Yahoo's security failings may embroil it with lawmakers.

Senator Mark Warner, D-VA, founder of the Senate Cybersecurity Caucus, said in a statement, "While I have repeatedly asked for briefings from Yahoo on the disclosure of its 2014 breach and have yet to receive a response, this most recent revelation warrants a separate follow up and I plan to press the company on why its cyber defenses have been so weak as to have compromised over a billion users."

An Uber self-driving car was just caught on camera running a red light in San Francisco. Today was the first day the cars were testing passenger pickup on the open road there.

Uber says the incident was the result of human error; the vehicle was being driven by a person at the time the video was taken, and there were no passengers inside.

A YouTube user known as Charles Rotter posted a video of the vehicle as it failed to stop at a red light, and drove through a crosswalk with a person clearly about to cross the street.

Here's the video:

youtube.com

Here's where the person who posted the video stops at the red light...

... and here's where what appears to be an Uber self-driving car blows right past it.

In his initial response to the video, Uber spokesperson Matt Kallman said in an email, "Safety is our top priority. These incidents have been reported and we are looking into what happened."

Meanwhile, the California DMV sent a letter to Uber today asking the company, which doesn't have a permit for autonomous vehicles, to cease the program.

Josh Edelson / AFP / Getty Images

Twitter spent more than a year developing an instant messaging app for emerging markets, but killed the product without launching it, BuzzFeed News has learned.

Built at Twitter’s Indian engineering center in Bengaluru, the standalone messaging app — which blended tweets and instant messages in a single interface — was conceived as a sort of new user on-ramp to Twitter proper. But the company killed it in September when it shut down its India engineering center.

Sources familiar with the messaging app’s development told BuzzFeed News that Twitter intended to identify “influencers” around certain topics — let’s say news or politics or sports — and encourage them to create groups of interest within the app.

Not only could users chat among themselves in these groups, they could also subscribe them to relevant Twitter accounts whose tweets would be pulled in automatically into these groups -- functionality that’s similar to what’s already possible in Slack channels.

“Twitter didn’t have that many active users in India,” one source who worked on the app’s development told BuzzFeed News. “So the idea was that if enough people used the instant messaging app, we could expose a lot of people to tweets without them even going to Twitter in the first place. Eventually, we hoped they would see the value of signing up for Twitter and directly following as many people as they wanted.”

Twitter refused to reveal the number of active users in India, but according to app analytics provider App Annie, the company had around 31 million monthly active users in the country on both Android and iPhone in November 2016.

The reason Twitter focused on instant messaging was simple — it was a concept that almost all users in emerging markets were already familiar with.

“Look, as a product, Twitter isn’t easy to figure out for most people,” said another source who directly worked on the app and said that none of his peers on Twitter’s engineering team in India were active Twitter users before they joined the company. “Everyone around us was hooked to WhatsApp and Facebook Messenger, and I think internally, there was some concern about how much people were engaging with those platforms versus ours. So instant messaging seemed like a natural choice to build something around.”

A project lead in Twitter’s Bengaluru office oversaw the instant messaging project but checked in frequently with a manager based in San Francisco for feedback, said the source.

The project, however, was ultimately killed since the versions of the app that Twitter tested out with users anonymously didn’t get good feedback. “It didn’t test out so well in the market surveys that we did with college students,” said the source. “And I think that’s one of the reasons why Twitter decided to shelve it.”

The app was killed in September after more than a year of development. Three weeks later, as a part of what Twitter described as a “normal business review” the company laid off its entire Bengaluru staff of 20 including nine software engineers. It also shut down its India engineering center, born out of the company’s $30 million acquisition of Indian startup ZipDial in January last year.

The move came a little more than a month before Twitter announced with its third quarter results that it would lay off 350 people -- 9% of its global workforce -- to cut costs. According to BuzzFeed News’ sources, even before Twitter made the layoffs official, the company’s focus had moved from user growth to revenue, and non-performing features and experiments — like the instant messaging app — were axed first.

Twitter declined comment on the messaging app, but a high-level source at Twitter India confirmed its development and fate. “We took a bet with this product but it didn’t work out,” the source, who requested anonymity, told BuzzFeed News. “After Twitter’s third quarter results, the company focused more on profitability versus growth, and that’s the reason why this app was put on the back-burner.”

“My personal opinion is that if Twitter’s stock hadn’t been falling like a stone, we could have got something else to work on, instead of being laid off like we were,” said another source. “Product testing happens all the time in tech companies and if something doesn’t work, generally, you move to the next iteration or the next thing. You don’t get fired.”

The objective of Twitter’s Indian engineering center was to build products for emerging markets, but multiple sources told BuzzFeed News that the instant messaging app was the only project that Twitter’s Indian engineering team worked on before being laid off.

“We briefly worked on Twitter Lite, a low-bandwidth version of Twitter for slow networks and cheap phones,” said a source. “But it was never released and we don’t know what happened to it.”

It’s worth noting that Twitter added some instant messaging features like read receipts, typing indicators and previews of web links to Direct Messages in the same month it killed off this messaging app.

Twitter’s India operations have been on shaky ground since last month. Less than a week after the company announced its global layoffs at the end of October, its India head, Rishi Jaitley, quit after a four-year stint. Just a couple of days later, Parminder Singh, Twitter’s managing director for Southeast Asia resigned too.

Both Jaitley and Singh came to Twitter from Google. Twitter closed offices in Germany and the Netherlands as part of the October layoffs, but according to Recode, most offices in Asia are still open.

When asked if emerging markets like India aren't a priority for Twitter anymore, the high-level source said they wouldn't go as far as to say that.

"However, there is a growing concern within the company that we shouldn't shut down," they said. "The priority now is to make money. We can revisit this app in the future if we think it's sustainable once again. Let's just say our ambitions have been tempered."

Stan Honda / AFP / Getty Images

Today, New York City Council Member Dan Garodnick will introduce a bill to create a public record of all sexual assault and harassment incidents that happen inside vehicles operated by New York City Taxi and Limousine Commission-licensed companies and owners (which include Uber, Lyft, and other ride-hail services).The bill is co-sponsored by Council Member Laurie Cumbo of Brooklyn, who is the chair of the Committee on Women's Issues.

The legislation was drafted in response to a BuzzFeed News investigation, which found that in the United States, Uber received five claims of rape and “fewer than” 170 claims of sexual assault directly related to an Uber ride as inbound tickets to its customer service database between December 2012 and August 2015. The investigation was triggered by Uber customer service database screenshots provided to BuzzFeed News; in one screenshot, a search query for “sexual assault” returns 6,160 Uber customer support tickets. A search for “rape” returns 5,827 individual tickets. Other variations of the terms yield similarly high returns: A search for “assaulted” shows 3,524 tickets, while “sexually assaulted” returns 382 results.

The legislation comes as rapes in cabs in New York City are increasing — 14 in 2015 compared with 10 in 2014. Currently, there's no requirement to make information available to the public about sexual assaults in the 143,674 TLC-licensed vehicles in New York City.

“We thought we should take the first step in here to try to combat this problem by measuring it,” Garodnick told BuzzFeed News. “A ride in a taxi should not be an invitation to a sexual assault and real data will allow us to take real action.”

The legislation will propose to collect the assault and harassment information from government entities, including the TLC, 311, and the NYPD. For now, individual taxi and ride-hail companies like Uber and Lyft will not have to report incidents, but Garodnick’s office is hopeful that the bill will expand to include them. The bill calls for the city to collect and aggregate this data annually, and to make the data publicly available and sortable by offense type (rape, verbal harassment, forced touching, to name a few) and vehicle type (for-hire cars like Uber and Lyft, yellow/green taxis, and livery cars).

“By making this information publicly available, it allows policy makers to take steps to clamp down on rape and sexual assault in cars. But it’ll also serve as a benchmark for all the stakeholders in the TLC space to see where they stand when it comes to preventing or being host to sexual harassment and assaults,” Garodnick said.

If passed swiftly, the Council Member’s office said the legislation could be enacted as early as this spring.

David Ramos / Getty Images

Facebook today announced several initiatives to help reduce the spread of fake news, and a major element involves giving fact checking organizations unprecedented prominence in the News Feed.

The largest social network in the world is partnering with organizations that have signed on to the International Fact-Checking Network fact-checkers’ code of principles to enable them to verify selected links being shared on Facebook and have those fact checks attached to the original link. This is the first time Facebook has given third parties special placement in the News Feed, which is the biggest single referrer of traffic to news websites in the United States, and a huge traffic driver in other parts of the world. This move comes after Facebook faced intense scrutiny for the spread of fake news and misinformation on its platform during the election.

“Symbolically, this is huge,” Alexios Mantzarlis, director of the ICFN, told BuzzFeed News when asked about the significance of this partnership for fact checkers.

He also cautioned that “we're going to have to wait and see how the solutions announced by Facebook work in practice and how they are scaled up worldwide to determine what significance this has for fact-checking and the battle against fake news.”

Adam Mosseri, the VP of product management for News Feed, told BuzzFeed News Facebook is not paying the checking organizations for their participation, but said their sites could benefit from the additional traffic that this new level of exposure could bring.

He also said this and the other new initiatives — which involve a tweak to the News Feed ranking algorithm, easier ways for users to report false content, and new ways to prevent scammers from making money from completely fake news — come in response to concerns about the spread of misinformation on Facebook.

Adam Mosseri

Steve Jennings / Getty Images

Facebook is initially focused on attacking “the worst of the worst” of fake news, according to Mosseri. He defined that as “clear hoaxes that are intentionally false and usually spread by spammers for financial gain.”

“Fake news is something we have were looking at before the last month or two, but I would say that the urgency around fake news has definitely increased given the feedback we received from the community,” Mosseri told BuzzFeed News.

How It Works

Here’s how the partnership works and how it will change how some links look in your News Feed: participating partners will have access to a special online queue that will show links Facebook determined may be suitable for a fact check. Links can end up in the queue because users have reported them as false, the viral nature of the link warrants a closer look, the source of the link is “masquerading as a news site,” or for example if a lot of comments say that the story is false or misleading. One or multiple partners can fact check a link.

If checkers rates the content of the link as false, the resulting fact check(s) will be attached to the link in News Feed, thereby alerting users to potential factual issues. So-called “disputed” links will also have their reach adversely affected in News Feed, according to Mosseri.

This is what a disputed piece of content will look like in News Feed:

Facebook

“[The fact checkers] can dispute an article and link to their explanation and then provide context on Facebook so people and the community can decide for themselves whether they want to trust an article or share it,” Mosseri said.

The queue of checkable links is already operational and at least some participating partners have begun submitting their articles to be attached to disputed content.

Concerns Over Bias

Conservatives in the United States have over the years criticised fact checking organizations for having a liberal bias. The owner of a large hyperpartisan conservative Facebook page and website recently told BuzzFeed News that he thinks Facebook as an organization also has a liberal bias.

Mosseri emphasized that this initial partnership is focused on purely fake news, and not on checking claims from politicians or wading into partisan disputes.

“Fake news means a lot of different things to a lot of different people, but we are specifically focused on the worst of the worst — clear intentional hoaxes,” he said.

Aaron Sharockman, the executive director of PolitiFact, told BuzzFeed News that his and other checking sites often face accusations of bias and said they focus on making their work as transparent as possible so readers can make up their own minds.

“I think that at some level you are never going to be able to satisfy certain critics, particularly very partisan ones,” he said. “At PolitiFact we’ve been doing this nine years and have published 13,000 fact checks and we have heard criticism of perceived bias from the left and from the right.”

He said he expected Facebook to take heat for aligning with fact checkers, but credited the company with taking steps to address the spread of fake news.

“They’re sticking their neck out a little bit and I think they’re gonna have to stick it out a little bit more because there are so many falsehoods flowing through Facebook feeds that Facebook is going to be very busy attacking these issues,” he said.

Meanwhile, Mantzarlis expects the partnership will result in a huge increase of applications to sign on to the International Fact Checking Network’s code.

“I imagine that inundated may be an understatement,” he said. “But if this decision by Facebook leads to a surge in genuine fact-checking projects, so much the better.”

In preparation for that deluge, and as a result of the new prominence checking organizations will receive in Facebook, the IFCN is reworking its vetting process.

“We’re now adapting the process to set up a vetting process and ensure compliance,” he said. “Aspiring signatories will have to go through that process. Existing signatories will also be vetted and if they don't meet the criteria, will be delisted.”

He also emphasized that signing onto the IFCN code is just the minimum requirement to be included in the Facebook checking project.

“First, it is important to note that Facebook decides which fact-checkers to include; the IFCN code is just the ‘minimum condition,’” he said in an email.

In response to a follow up question about how organizations will be approved for the third-party fact checking program, a Facebook spokesperson said that participants “must be signatories” of the IFCN code. As of this writing, six US-based organizations have signed on: Associated Press, PolitiFact, FactCheck.org, Snopes, The Washington Post, and ABC News. The latter was only added as a signatory two days ago, and AP joined the day of the announcement.

Mantzarlis said ABC News applied early last week. Its website lists two stories for its Fact Or Fake feature. AP has been publishing fact checks for several years.

Caeden

Sona, a stress-tracking bracelet marketed under the premise that it could calm its wearers, has ended up stressing out customers whose gadgets quickly stopped working or didn’t arrive on schedule. And although the wearable’s maker, Caeden, is taking steps this week to refund people or finally ship their orders, many customers complain that the startup failed to acknowledge their frustration for weeks, even months.

Caeden was supposed to ship thousands of Sona bracelets to people starting in late August, followed by a second batch in late November. The stylish wearable, which BuzzFeed News wrote about when the New York City startup started taking pre-orders last fall, was designed to measure wearers’ heart rate variability. Paired with an app, it intends to help people manage their focus, stress, and anxiety.

But unhappy buyers told BuzzFeed News that their Sonas — which cost $129 to pre-order and $199 at full value — stopped working soon after arrival. “Imagine my thrill when, within 48 hours, the bracelet stopped working,” one wrote in an email. “Completely refused to communicate with the app.” Some of them said they didn’t receive refunds even when they returned the bracelets. Others said that their deliveries had been repeatedly delayed without explanation.

And customers in both camps told BuzzFeed News that Caeden had ignored their requests for help, as indicated by dozens of complaints on Caeden’s Facebook and Twitter accounts. The first tweets about malfunctions started appearing around late August and early September.

Screenshot of Caeden's Facebook page, taken Dec. 14.

Caeden is now scrambling to make amends. On Wednesday, the startup emailed customers to tell them that, depending on their situation, their refunds would be processed or their orders would arrive after the holidays. It posted an apology and explanation on its Facebook page Thursday. And on Friday, it plans to answer questions through a virtual town hall on YouTube Live.

CEO Nora Levinson told BuzzFeed News that the devices malfunctioned due to a coding error in a part from a third-party supplier, and the bug caused the bracelet to stop syncing with the app. When the bug began popping up, it took the 10-person staff until October to figure out what was happening, Levinson said.

At that point, Caeden stopped production on the new pre-orders and stopped accepting new orders. During that time, customer service slowed down, Levinson said.

“Because this was such a widespread issue and so difficult to figure out, we essentially very quickly became overwhelmed with the response,” Levinson said.

The startup took such a financial hit as a result of the bug, Levinson said, that it laid off half its 10-person staff and tightened salaries. She said that the problems with the outside supplier have been fixed.

“Our number-one priority as a company is resolving this with everybody and being able to move forward and trying to help people get this situation resolved,” she said. “We apologize for the delay and frustration that it’s taken in order to get to this point. I know it’s not been ideal in terms of a lot of people’s experience.”

Palantir CEO Alex Karp at Trump Tower on Wednesday.

Andrew Kelly / Reuters

The market for stock in Palantir Technologies, the Silicon Valley data analysis company that does much of its business with the federal government, was in the doldrums for much of 2016. But since the election of Donald Trump — whose closest tech adviser is Palantir chairman Peter Thiel — demand seems to have picked up, one investor and middleman in that market has told BuzzFeed News.

“Since the election, we’ve seen an uptick in interest in owning Palantir shares," said Jared Carmel, a managing partner of Manhattan Venture Partners, a boutique investment bank and fund manager that arranges trades in startup shares, and which owns a stake in Palantir.

Palantir's elevated stature was on display this week when its CEO, Alex Karp, joined the tech industry's biggest players in a pilgrimage to Trump Tower. Others in the room were top brass from Apple, Google, Microsoft, Amazon, and Facebook — public companies each valued in the hundreds of billions of dollars. Karp's Palantir, which is privately held, has a relatively small valuation of $20 billion.

Given Palantir's roster of customers — which includes the FBI, the CIA, the Defense Intelligence Agency, the Securities and Exchange Commission, Immigration and Customs Enforcement, the military's Special Operations Command and other federal agencies — this face time with Trump and his closes advisors was particularly notable.

Palantir, Carmel said, now appears to be in a strong position to expand its work with the federal government. Trump "is nepotism 101. Any additional contracts that I imagine they could get, they will get."

For most of 2016, the market for Palantir stock was "dead," Carmel said. Former employees of Palantir struggled to cash out of the stock options that formed a major part of their pay packages, BuzzFeed News reported in October. Carmel said a BuzzFeed News article in May, which revealed Palantir had lost blue-chip clients and was struggling to stem staff departures, had contributed to the market lull. But things apparently turned around, at least a little bit, in early November.

Investors including hedge funds, family offices, and venture capital funds have recently inquired about buying Palantir shares, Carmel said. Manhattan Venture Partners helps investors buy shares of tech startups that it believes have a good chance of going public soon, and it manages funds that invest in those companies.

Backing off his longtime reluctance to plan for an IPO, Karp said in October that he was "now positioning the company so we could go public. I'm not saying we will go public, but it's a possibility."

A Palantir spokesperson did not immediately respond to a request for comment.

It's not just Manhattan Venture Partners that is getting interest in Palantir stock. This week, in a private Facebook group for Palantir alumni, one former employee said he had learned that EquityZen, a firm that arranges small sales of private company shares, was fielding investor demand for Palantir shares.

"I'm sure some of you guys got an email but EquityZen has some investor interest and has asked anyone that would like to sell to fill out their profile with # of shares, desired price, etc," the former employee wrote. Contacted by BuzzFeed News, EquityZen declined to comment.

"I would post more details," the former Palantir employee wrote, "but in case there's a BuzzFeed troll (hi William), DM me if you didn't get the email and I can gladly send over."

AP/Richard Vogel

Federal law now guarantees the right to leave a bad Yelp review, after President Obama signed the Consumer Review Fairness Act on Wednesday.

The law, which passed the Senate in November with bipartisan support, comes after years of legal efforts to overturn so-called "gag clauses" in contracts, which consumer advocates say are used by businesses to stifle free speech. The clauses, which have been included in the fine print of sign-up forms and purchase agreements, forbid public criticism of companies by their customers.

Such clauses are banned under the new law, which also prohibits businesses from imposing a penalty or fee on a client if they write a bad review. Consumer advocacy groups and other supporters of the law are celebrating its passing.

Yelp, whose users have been sued by companies over negative reviews, said the Act "gives Americans nationwide new guaranteed legal protections when it comes to sharing these honest, first-hand experiences."

Paul Levy, an attorney with Public Citizen who has represented a number of consumers who have been sued over gag clauses, said businesses have used the legal fine print to "suppress truthful criticism and skew the marketplace of ideas by allowing only positive reviews and suppressing negative ones."

"Non-disparagement clauses not only prevent consumers from expressing themselves but also deprive prospective customers of balanced and truthful information on which they can base their decisions about where to do business," he said. "What’s worse, most consumers are unaware when they have agreed to a non-disparagement clause curtailing their First Amendment rights."

Levy represented one of the first consumers to be sued by a company over a negative review. A company called KlearGear claimed a couple in Utah owed a debt for violating its terms of service by writing the review. The couple sued KlearGear in 2012 arguing they didn't owe any debt to the company and received a default judgement including over $300,000 in damages when the company never appeared in court.

In another case in 2014, a Wisconsin woman filed a lawsuit against Accessory Outlet, an online mobile retailer, after they demanded she pay $250 for violating its terms of sale. The company never appeared in court to defend its claims and the court granted the woman a default judgment in 2015, including triple the damages and attorneys fees.

California made gag clauses in contracts illegal in in 2015. Any business that violates the law will be fined up to $2,500 on its first violation and $5,000 for following violations. Consumers can receive up to $10,000 if they can prove the business acted recklessly in violating the law.

LINK: Yelp’s Warning: This Dentist Might Sue You For Posting A Negative Review

LINK: Someone Is Filing Wild Lawsuits Aimed At Negative Online Comments

LINK: Gag Clause 2.0: Companies Tell Customers To Keep Quiet If They Want To Cancel

LINK: Court Upholds The Right To Leave A Bad Yelp Review

Microsoft CEO Satya Nadella enters Trump Tower.

Andrew Kelly / Reuters

In response to questions from BuzzFeed News, Microsoft spokesperson Frank X. Shaw clarified his company's position on the use of customer data. “We’ve been clear about our values. We oppose discrimination and we wouldn’t do any work to build a registry of Muslim Americans," said Shaw.

The company's statement comes a day after its CEO, Satya Nadella, attended a tech summit hosted by President-elect Donald Trump, and at a moment many when tech leaders are under increasing pressure — from both their own employees and the public — to explain how their companies would respond to government requests from the incoming administration, including being asked to build a Muslim registry.

Microsoft initially declined to comment on hypotheticals when BuzzFeed News asked earlier this week, pointing instead to a blog post by Microsoft's president and chief legal officer Brad Smith written the day after the presidential election. The company goes into more detail about its principles and policies related to law enforcement and national security requests in its digital trust report.

Facebook also initially would not comment on the record about a Muslim registry, but later told BuzzFeed News: "No one has asked us to build a Muslim registry, and of course we would not do so." This week employees from Microsoft, Facebook, and other tech companies signed a pledge not to help build a Muslim registry and to fight any illegal or unethical data practices.

During his campaign, Trump repeatedly stated that he wanted to build a database of Syrian refugees and made bigoted remarks about Muslims. When asked about a Muslim registry, Trump would not deny that he wanted one, but said his priority was the database of Syrian refugees.

Microsoft has sued the government four times recently to challenge requests for information and Microsoft's ability to be transparent about those requests. One lawsuit challenged a nondisclosure order attached to a national security letter. The case resulted in the government withdrawing the letter. In 2013, documents leaked by Edward Snowden showed that Microsoft collaborated with US intelligence services to allow access to communication from users as part of a top secret program called PRISM.

Yesterday's meeting, which was organized by Trump's billionaire backer, venture capitalist Peter Thiel, and Trump's son-in-law, Jared Kushner, covered jobs, technology's impact on the economy, and possibly repatriation of offshore cash.

BuzzFeed News; Getty

In early July, a website made to look like a real news organization published what would prove to be one of the biggest fake news hits of the US election.

WTOE5News.com was barely two weeks old when it published the hoax story, “Pope Francis Shocks World, Endorses Donald Trump for President, Releases Statement.”

It improbably quoted the pope as saying that the FBI’s inability to prosecute Hillary Clinton for her emails led him to endorse Trump. The story also falsely declared that “news outlets around the world” were reporting the endorsement.

To date, the hoax has registered over 100,000 comments, shares, and reactions on Facebook, according to data from BuzzSumo. But the hoax did even better on Facebook when, in late September, the website Ending the Fed published a fake story with the exact same headline. That post has earned close to 1 million Facebook engagements and was the single biggest fake news hit of the election, according to a previous analysis by BuzzFeed News.

Ending the Fed’s copycat hoax has gone on to capture additional attention as concerns about fake news during the election have intensified. That hoax has been pointed to again and again. Yet little has been said — and is known — about the site that originated this massive fake news hit.

WTOE 5 News is no longer online, and its owner has never been identified. It has largely been overlooked as a player in the world of fake news, a flash in the ersatz pan. But a BuzzFeed News investigation has found that the site is part of one of the world’s most unique and ambitious fake news operations — a network of at least 43 websites that together have published more than 750 fake news articles.

Many of the fake stories identified by BuzzFeed News followed the exact same pattern: They falsely reported that a big celebrity was moving to a specific town or community.

None of the sites list an owner or company, and all of their domain registration records are private. Not surprisingly, the person behind one of the biggest fake news scams of all time prefers to remain hidden. To try and identify the owner, BuzzFeed News followed a trail from the fake sites to a group of now-defunct websites about topics including nurses and photos of Jennifer Love Hewitt, and eventually to a house in the small city of Atascadero, California, where a man named Justin Smithson resides.

This is the story of one of the world’s largest and most unique fake news empires, and how it gave birth to what became the iconic hoax of the 2016 election.

Steve Arrison thought his job was about get a lot easier.

Arrison is CEO of the Hot Springs Convention & Visitors Bureau for the town of Hot Springs, Arkansas. Just under 40,000 people live there, but it welcomes an estimated 3 million tourists a year. Arrison’s job is to bring in as many of those visitors as possible. In early March, a news story had the town thinking they’d landed a major VIP and tourist draw.

“Clint Eastwood is Moving to Hot Springs, Arkansas” read the headline. It was being shared like crazy by locals. Even state media was starting to call about it. But as soon as Arrison clicked the link and started reading, he realized something was off.

“I went to the web and saw that same site said Katy Perry was moving to some place in Texas, so immediately a red flag went up,” Arrison told BuzzFeed News. “A lot of people wasted a lot of time on something that was blatantly false. Information moves so quickly it takes a while for the truth to catch up to the falsehoods.”

Beginning in late February, there was a sudden onslaught of news stories on Facebook that all claimed one celebrity after another was moving to an unexpected place. The stories had headlines such as “Johnny Depp Explains Why He’s Moving to Pittsburgh, Pennsylvania,” “Matthew McConaughey Explains Why He’s Moving to Greenville, South Carolina,” and “Demi Lovato Moving to Perris, California.” The list of celebrities in the hoaxes began to expand and soon Eminem, Ryan Gosling, Rihanna, Jim Carrey, Samuel L. Jackson, Justin Bieber, and Brad Pitt were moving. As the days and weeks went on, the locations also began to move farther away on the map: Vin Diesel was moving to Saskatoon, Saskatchewan. Leonardo DiCaprio was moving to Guildford, Surrey, UK.

These stories appeared on sites with legitimate-sounding news domains such as kspm33.com, mckenziepost.com, ky6news.com, and km8news.com. The design of each site was strikingly similar — often just the colors and the name were changed. The text, too, was a simple copy-and-paste effort; just the celebrity’s name and location were changed from one story to the next. It quickly proved to be an effective strategy as the stories began to spread across the internet.

One morning in early March, a staffer for venerable rumor-checking website Snopes logged on to the site’s main email account to read the tips and questions that had flowed in overnight. She saw something unusual: Roughly 25 emails that all asked about reports of different celebrities moving to different towns. The same rumor was being repeated again and again — but with different celebrities and different locations.

Kim LaCapria, a writer for Snopes, told BuzzFeed News they were initially puzzled at how “scattered” the rumors were.

“That spate of rumors was scattered across locations and individuals and basically was diffused to the point where the average social media user would be less apt to ‘see’ the rumor,” she said. “Particularly because they’re seeing it from their vantage point, and likely to think, ‘Well, who’s going to take the time to lie or make up a hoax about my tiny neighborhood?’”

The local focus ensured the hoaxes spread in targeted clusters, and meant someone in, say, Saskatoon had no idea that the same trick was being played on people in Pittsburgh.

Snopes eventually wrote a single post to warn people that no, that big celebrity was almost certainly not moving to their town. One simple clue was that the about page for each site had the exact same disclosure telling people that they were reading a “a fantasy news site.” But the stories kept coming and people kept getting fooled.

LaCapria said in an email that “the rumor was so atypically benign (clearly not a potential lie to advance any agenda or belief) that it again just sort of managed to infest tons of local areas.”

BuzzFeed News identified 342 celebrity-moving hoaxes spread across multiple, connected sites. As the scam scaled and hit more communities, local news websites began publishing debunkings to keep locals from falling for the hoaxes. News outlets in Texas, Maine, Illinois, South Carolina, British Columbia, and many other places tried to stop the stories from spreading.

Meanwhile, the growing network of fake sites began pumping out new variations. Male celebrities suddenly had very complimentary things to say about the women in specific towns. Celebrities began getting flat tires in obscure places, naming somewhat obscure locales as their favorite vacation spots, and buying homes in unexpected communities. One of the strangest variations was a series of at least 11 hoax stories claiming that Justin Bieber was building megachurches in places such as Spokane, Washington.

Another variation moved away from specific celebrities and instead falsely claimed that big film sequels were being shot in different locations: Father of the Bride III, a Harry Potter spinoff, a new Star Wars Film, a new Pretty Woman. They were all suddenly in production and coming to your town. BuzzFeed News identified 152 fake news articles claiming that major film sequels were being filmed in different locations.

The hoaxes combine two important elements that help them spread: location and name recognition. Many people saw a story about their town that also included a celebrity or major film and couldn’t help but share the link on Facebook, thereby driving traffic to sites littered with ads.

An analysis of the sites' content, done using the BuzzSumo content analysis tool, found that while these hoaxes never went hugely viral, they managed to consistently generate tens of thousands of shares, comments, and reactions on Facebook, which likely led to strong traffic and revenue.

BuzzFed News began internally calling this the “local viral scam” because the fake stories went viral on a local basis but never bubbled up higher than that.

LaCapria of Snopes admitted she has grudging respect for the way this approach combined someone or something famous with different locations to spark social shares.

“It’s a very unique thing, the execution was so different,” she said. “It only works if you plug in a relevant celebrity to the area. It was really well targeted.”

It was also proving lucrative enough to keep expanding. BuzzFeed News was only able to identify the existence of one site, McenziePost.com, in February. But at the end of April there were more than 20 nearly identical fake news sites. By June at least 27 were in operation. That was also the month WTOE 5 News was registered. A few weeks later, in a departure from the template of previous fake news articles, that site published the pope endorsing Trump story. It earned over 100,000 shares, comments, and reactions on Facebook, making it the network’s fourth biggest hit on the platform, according to an analysis done using BuzzSumo.

Not long after, on July 19, yet another newly registered site published the exact same story with a twist: This time the pope endorsed Hillary Clinton. (The site that hosted it, KYPO6, it is no longer online but used the same template and language as other sites in the network.)

That story performed even better, generating over 200,000 shares, comments, and reactions on Facebook, according to BuzzSumo. (Both stories were published before the party conventions.)

A few other websites picked up the fake papal story fake story in July, but none saw significant engagement. The Clinton story, in spite of being a bigger hit on Facebook, did not inspire copycats. For the most part, those two hoaxes came and went, and the network of fake sites went back to their core focus of churning out local viral stories.

Then, as the election kicked into high gear in September, Ending the Fed published a hoax story with the exact same headline and scored a massive hit. (Ending the Fed is an anonymously run website that previously attracted attention when Facebook promoted one of its false stories as a trending topic.)

Perhaps inspired by the success of that copycat story, the papal hoax story made yet another appearance on one of the local viral fake sites, Channel16News.com, in mid-October. But it failed to gain much traction on Facebook.

Since then, the network of fake sites has continued to grow, reaching at least 43 sites. The most recently registered domain identified by BuzzFeed News, DailyNews5.com, was bought on Nov. 3. That’s a few days prior to the US election — and before alarm bells about fake news began going off in the media and among political leaders.

But rather than completely shut down the operation and move on in order to avoid attention, the sites continue to push out new hoaxes. A fake story about a new Star Wars film being shot in Roswell, New Mexico, went online just this week, according to data in BuzzSumo. Whoever is running the sites appears to be unconcerned by the outrage over fake news. Or perhaps the sites were too lucrative to shut down? We wanted answers. So BuzzFeed News decided to follow the money.

Though some are now offline, nearly all of the sites in this fake news network contained the same Google AdSense ID in their source code. This means the money earned each month from ads placed on the sites is going into a single account. That ID has also appeared on other websites over the years, from TehYouTube.com and NursingJobResources.com to JenniferLoveHewittPics.com and KimKardashianPics.com. Google does not publicly list the owners of AdSense IDs, but the aforementioned sites were publicly registered by the same person: Justin Smithson, who lived in Missouri at the time, according to historical registration data from Whoisology and DomainTools. The Justin Smithson who resided at that Missouri address now lives in Atascadero, California, according to records obtained in a LexisNexis search. (Go here to view a collection of documents linking Smithson and the network of fake sites to the AdSense ID.)

On the left is a list of some of the websites connected via a single AdSense ID. On the right is a public domain registration record for JenniferLoveHewittPics.com and other sites registered by the same Justin Smithson.

CuteStat / DomainTools / Google

Smithson is a pilot in his thirties who works for a charter airline in San Luis Obispo. Over the years he has registered and launched many websites. At some point, it appears that Smithson signed up for Google AdSense in order to place ads on his websites. As he launched new sites, Smithson continued to use that same AdSense code, leaving a trail that eventually connected him to the network of fake news sites. (There is of course a chance that Smithson registered those sites on behalf of someone else, who then added AdSense and went on to later launch the network of fake sites. Or that Smithson handed over his AdSense account to someone else.)

In addition to being an experienced webmaster, public Facebook comments from Smithson also indicate he has above-average knowledge of computers and the internet. These two public Facebook comments (which were later deleted) show him talking about the frame rate output of his computer’s graphic card:

Facebook

Domain registration records also show that Smithson continues to register new domains using the same email address he used for JenniferLoveHewittPics.com and others. At the end of November, for example, he registered FoxBusiness.xyz. In September he bought AlexJones.xyz. As with the fake news sites, those domains now have their ownership information hidden.

In early October, BuzzFeed News contacted Smithson by email to request an interview about his fake news websites. “I'm guessing this is some kind of spam/scam but if not... You've got the wrong person,” he replied.

Smithson did not reply to subsequent emails, including one on Oct. 5 requesting a meeting at noon. On Oct. 6, BuzzFeed News visited his place of work and his Atascadero home, where we found him on the front porch with another man. When they men saw us approaching with a camera in hand, they went inside.

We called out to say we wanted to talk to Justin, and one of the men yelled from a window that there wasn’t anyone there by that name. He threatened to break our camera, so we left.

Smithson did not reply to four subsequent emails, texts, and phone calls. Nor did he reply this week when BuzzFeed News sent him detailed questions and shared the evidence that points to him as the owner of perhaps the world’s biggest fake news operation. (Aside from any state-sponsored efforts, that is.)

As of this writing, at least 20 of these websites are still online. Their stories are still being shared, and they still publish new fake news stories that fool people and earn money for that same AdSense account. LaCapria of Snopes says she still sees new variations of the local viral scam from time to time.

“It still seems to be living on — but most people have gotten past asking us if it’s true that a celebrity is moving to their town," she said.

Mc3 Bryan Mai / U.S. Navy

Killer robots, and whether to ban them, will be on the menu for international talks next year at the United Nations. The decision was approved on Friday by a panel that included the US and China.

The weapons under consideration, prosaically known as “lethal autonomous weapons systems,” include robots that can find a target and fire at will without a person making that decision.

Sixty-five non-profit groups united under the “Campaign to Stop Killer Robots” cause have been calling for a ban on such weapons for over three years.

The UN agreement represents a first-time handoff of the debate to governments, which include some countries who have spent billions of dollars developing such weapons.

The decision was made at the UN Convention on Certain Conventional Weapons, the group that bans or regulates the use of very new or very lethal weapons. In the past, the CCW has resolved to ban landmines, incendiaries like napalm, and blinding lasers.

Following protocol, nations agreed to form an expert group who will next year debate how killer robots could be regulated.

“It puts the power into their hand as governments,” Mary Wareham, advocacy director of the Arms Division at Human Rights Watch and coordinator of the Campaign To Stop Killer Robots, told BuzzFeed News. “Until now, it’s been a series of PowerPoints by experts and it’s governments asking them questions. And this is where they go home and go do their homework.”

Earlier this month, nine members of Congress led by Massachusetts Rep. James McGovern addressed a letter to the Secretaries of State and Defense supporting continuing the expert talks, and recommending that the US delegation support a four-week meeting next year.

“As this military technology continues to advance, we need to take a hard look at the risks it poses,” McGovern said in a statement emailed to BuzzFeed News.

Mc2 Antonio P. Turretto Ramos / U.S. Navy

Experts agree that the campaign has had swift success calling attention of the international community to a largely futuristic technology.

“If you look at the last review conference in 2011, not a single state party even mentioned autonomous weapons,” Southern Methodist University law professor Chris Jenks who spent 20 years with the US Army, then worked on international law at the Pentagon, told BuzzFeed News. “Here we are 5 years later, and it is the focus.”

But despite this success, Jenks and others believe that an outright ban on killer robots is unlikely to be the result of next year’s talks.

“I think it’s going to be a really slow process and we’re not likely to see the military powerhouses signing on to any treaty ban,” Rebecca Crootof, a lecturer at Yale Law School, told BuzzFeed News.

So far, 19 countries have called for a ban (three of those spoke at this week’s deliberations), but Crootof pointed out that the list did not include many countries — like the US, China and Russia — that have already spent billions building those weapons. The US has focused its efforts on self-driving trucks and helicopters, and drone swarms that coordinate flight among themselves.

Crootof also argues that a more reasonable approach would be to regulate how they are used, as the expert discussions could do, because of how unique this class of weapon is.

“Autonomous weapons systems are just fundamentally unlike anything else we’ve regulated before,” she said.

Others are cautiously optimistic that automated decision making, ironically, could lead to more humane warfare.

“If weapon systems can be developed which are more discriminating and more accurate than current ones, not only should they be developed, I think there is a moral imperative to do so,” Jenks said.

“I never encountered someone who felt less bad about being shot because they were shot by a human — there is being shot and there is not being shot.”

131109-N-ZZ999-176

Mcsn Hilkowski / U.S. Navy

LINK: How To Save Mankind From A New Breed Of Killer Robot

LINK: The U.S. Military Is Betting On “Smart” Drones — Lots And Lots Of Them

LINK: Drones Could Become Flying Peeping Toms, Privacy Experts Warn

Facebook Messenger is on a roll, pushing out new features in rapid succession as 2016 comes to a close.

A week after releasing a new, Snapchat-style camera, the wildly popular messaging app is introducing group video chat, its most requested feature. Group video chat on Messenger will be available globally, both on iOS and Android, and will start rolling out Monday.

Messenger's impressive run of new feature releases (it's also experimenting with AI-powered contextual suggestions) is evidence Facebook is unwilling to simply let it ride the coattails of the 1.79 billion user core app it's inherently tied to. Instead, as Messenger ships fun new features, Facebook is proving the wisdom of spinning out the app in 2014, a move initially met with user backlash.

Group video chat on Messenger will support up to six videos at a time, but up to 50 people can watch and, if they so choose, join via voice, stickers and the like. A video icon will now appear at the top right of group conversations, and tapping on it will alert group members and give them the choice to join.

Group video chat is having a bit of a moment right now. Houseparty, a group video chat app by the creators of Meerkat, just raised $50 billion. And Fam, an iMessage group video app, seems to be taking off as well.

One big question about Messenger remains: how will it make money? At this point, the app is very popular — claiming over 1 billion users — but it's not contributing meaningfully to Facebook's bottom line. Facebook is anticipating a slowdown in ad load growth, so it may soon need to make a revenue push inside Messenger, something that lead to even more changes. But in the meantime, enjoy your video chats.