Brain-computer interfaces are developing faster than the policy debate around them
Illustration by James Bareham / The Verge
A few days ago, Facebook disentangled itself from a nettlesome investigation by the Federal Trade Commission into how the company violated users’ privacy. And then, with that matter now squarely behind it, Facebook on Tuesday stepped forward to share some information about its effort to read our minds.
Two years after the company announced its mind-reading initiative, Facebook has an update to share. The company sponsored an experiment conducted by researchers at the University of California San Francisco in which they built an interface for decoding spoken dialogue from brain signals. The results were published today in Nature Communication.
The work itself is fascinating, as you might expect from the subject matter. Brain-computer interfaces aren’t new, but the existing ones aren’t particularly efficient — particularly the ones that don’t involve drilling into your skill. Facebook’s approach relies on high-density electrocorticography, aka ECoG, which implants sensors on the brain and uses them to record brain activity.
And its most recent research apparently showed promise, Adi Robertson reports:
If participants heard someone ask “Which musical instrument do you like listening to,” for example, they’d respond with one of several options like “violin” or “drums” while their brain activity was recorded. The system would guess when they were asking a question and when they were answering it, then guess the content of both speech events. The predictions were shaped by prior context — so once the system determined which question subjects were hearing, it would narrow the set of likely answers. The system could produce results with 61 to 76 percent accuracy, compared with the 7 to 20 percent accuracy expected by chance.
“Here we show the value of decoding both sides of a conversation — both the questions someone hears and what they say in response,” said lead author and UCSF neurosurgery professor Edward Chang, in a statement. But Chang noted that this system only recognizes a very limited set of words so far; participants were only asked nine questions with 24 total answer options. The study’s subjects — who were being prepped for epilepsy surgery — used highly invasive implants. And they were speaking answers aloud, not simply thinking them.
If successful, the work will have important clinical applications — it could help patients to communicate who have lost the ability to speak, for example. Facebook hopes the technology has a broader use — enabling what former Facebook crazy-project chief Regina Dugan once called a “brain click.” Allow people to click through dialog boxes with their minds, she told us in 2017, and you create lots of interesting new possibilities for augmented and virtual reality.
That goal remains very far away. But that seems like a good time to ask whether any of this work should, you know, be done in the first place. Antonio Regalado’s piece on the Facebook experiment gets at why:
“To me the brain is the one safe place for freedom of thought, of fantasies, and for dissent,” says Nita Farahany, a professor at Duke University who specializes in neuro-ethics. “We’re getting close to crossing the final frontier of privacy in the absence of any protections whatsoever.”
Facebook, for its part, included a section on ethics in its blog post on the subject, quoting Mark Chevillet, director of the brain-computer interface (BCI) research program at Facebook Reality Labs:
“We can’t anticipate or solve all of the ethical issues associated with this technology on our own,” Chevillet says. “What we can do is recognize when the technology has advanced beyond what people know is possible, and make sure that information is delivered back to the community. Neuroethical design is one of our program’s key pillars — we want to be transparent about what we’re working on so that people can tell us their concerns about this technology.”
It has also pledged to have its research governed by an ethics board.
Of course, at this point, even invasive technology can barely distinguish between a speaker saying “fertilizer” versus one saying “synthesizer.” But it’s in the nature of these technologies to improve exponentially, often away from public view, and to mature before any real public conversation about them can take place.
And so it’s worth noting that Facebook hasn’t ruled out using brain activity for advertising purposes at some point in the future. In some ways, it feels like the logical conclusion of an advertising monolith. Its whole business is predicated on reading your mind however it can, whether by getting you to share all of your demographic data in a profile or by reading your brainwaves through a cap on your skull. In some ways, it would be weirder if Facebook didn’t seek to use your brain activity for advertising purposes.
Brain-computer interfaces would seem to have a lot of promise for medical uses — something that Elon Musk, who is also exploring the technology through his company Neuralink, is currently focused on. But they would also seem to carry with them a great risk of anti-democratic surveillance.
A futuristic headset that reads our minds to let us click through dialog prompts is all well and good. But I worry about how it might be used should the technology get much better than that. And that’s probably a conversation we want to start before Facebook and other companies make too much progress.
Democracy
New bill would ban autoplay videos and endless scrolling
Here’s a fascinating if likely doomed bill from Sen. Josh Hawley (R-MO) that would ban many popular growth hacks and engagement mechanics. From Makena Kelly:
Hawley’s Social Media Addiction Reduction Technology Act, or the SMART Act, would ban these features that work to keep users on platforms longer, along with others, like Snapstreaks, that incentivize the continued use of these products. If approved, the Federal Trade Commission and Health and Human Services could create similar rules that would expire after three years unless Congress codified them into law.
“Big tech has embraced a business model of addiction,” Hawley said. “Too much of the ‘innovation’ in this space is designed not to create better products, but to capture more attention by using psychological tricks that make it difficult to look away.”
TikTok is exhibit A in Facebook’s “we’re no monopoly” case
TikTok’s momentary success has offered Facebook a fig leaf of evidence that Facebook is not a monopoly, David McCabe reports. Counterpoint: do you knowhow many Facebook ads TikTok had to buy to get this far?
People familiar with Facebook’s thinking on competition issues said Facebook views TikTok’s rapid growth as one key example in its broader antitrust defense.
The company particularly sees it as evidence that Facebook’s reach has not limited the barriers to entry for new startup social products, one of the people said.
Canada’s New Partisan Media Are Poised To Ride A Facebook Wave To Election Day
A big team from BuzzFeed chronicled the rise of hyper-partisan media in Canada on Facebook:
North99 is part of a new wave of advocacy media — both on the left and right — that is reaching more Canadians than ever before, thanks to savvy approaches to Facebook. These pages and associated websites are generating millions of shares, reactions, and comments, and at times eclipsing the country’s biggest newspapers and broadcasters on the platform, according to a new analysis by BuzzFeed News and the Toronto Star.
But the lines between traditional news reporting — with clear editorial standards and accountability — and these new media players can be difficult to discern, especially when you’re distractedly scrolling through your newsfeed, according to experts.
Here’s where Facebook’s record $5 billion fine goes
Turns out it just goes into the U.S. Treasury! Which sounds like it could be cool until you remember that the national debt is $22 trillion.
Trump’s Baltimore Tweets Don’t Violate Company’s New Rules
An emerging genre of story is “Twitter does nothing about the president’s tweet.” Most recently, Twitter did nothing about some tweets involving a black member of Congress and his district.
China’s Beijing Kunlun to revisit Grindr IPO
In March, the US government took action suggesting that it considered gay hookup app Grindr a national security issue. But in a surprise twist, it reversed course this week, report Meg Shen and Echo Wang:
Chinese gaming company Beijing Kunlun Tech Co Ltd said on Monday it would revive plans for an initial public offering (IPO) of popular gay dating app Grindr Inc, after a U.S. national security panel dropped its opposition to the plan.
Kunlun said in May it had agreed to a request by the Committee on Foreign Investment in the United States (CFIUS) to sell Grindr, setting a June 2020 deadline to do so and putting preparations for an IPO of Grindr on hold.
Elsewhere
Gilroy gunman promoted white supremacist treatise on Instagram
Ugh:
Shortly before Sunday’s mass shooting in Gilroy, the 19-year-old who committed the crime used his Instagram account to promote a book widely considered a manifesto of white supremacy, offering insight into the still-unknown motivation of the gunman. […]
In an Instagram post written about an hour before the rampage, [the gunman] encouraged people to read “Might Is Right or the Survival of the Fittest.” The fringe book, originally published in 1890 and since re-released by a handful of small publishing houses, is a call to action against the alleged tyrannies of government and organized religion. The 96-page work encourages the “strong” to rise up over the “weak.”
Apple joins Google, Facebook, and Twitter in data-sharing project
Russell Brandom reports on a rare positive, pro-competition data sharing initiative between tech giants and their much smaller competitors:
Just over a year after its official launch, the Data Transfer Project is announcing a new set of partners and features. Today, Apple announced that it will be joining the project, developing interoperable systems to bring data in and out of iCloud. A number of alternative social networks have also joined the project, with Tim Berners-Lee’s Solid project enabling the import and export of contacts, and Mastodon allowing for the import and export of posts.
An open-source project aimed at making it easier to transfer data from one service to another, the Data Transfer Project has mostly consisted of back-end coding to make data export tools like Google Takeout and Facebook’s Access Your Information tool compatible with each other. Right now, those tools let you download data directly to your hard drive, but the hope is that the project’s code could allow the data to be ported directly to another service. That would allow you to send all of your Facebook photos to a Google Photos account, for instance, with no intermediate step and a lesser possibility of leaked data.
Facebook design flaw let thousands of kids join chats with unauthorized users
As someone who has found Messenger Kids icky from the start, I read this one with an extremely smug expression:
Facebook’s Messenger Kids app is built around a simple premise: children shouldn’t be able to talk to users who haven’t been approved by their parents. But a design flaw allowed users to sidestep that protection through the group chat system, allowing children to enter group chats with unapproved strangers.
For the past week, Facebook has been quietly closing down those group chats and alerting users, but has not made any public statements disclosing the issue.
Facebook Watch Draws More Advertisers
Sahil Patel finds signs of life on Facebook’s video platform:
Advertisers have warmed up to Watch somewhat as time has passed, buyers said.
One senior media buyer, who oversees more than $350 million in marketer ad spending each year, said he plans to spend at least $1 million this year on ads inside Watch. That’s not an enormous sum, but he didn’t advertise on Watch at all last year. It will be part of a bigger deal the buyer has signed with Facebook to spend more than $10 million on Facebook video ads in 2019, up 50%, the buyer said.
Twitter Adds 5 Million Users; Sales Top Estimates
It was very odd on vacation to learn that Twitter is doing … reasonably well? Like, as a company?
Second-quarter revenue came in at $841 million, up 18% from the same quarter last year and higher than the $829 million analysts estimated. The company added 5 million daily users in the period, bringing the total to 139 million. That was the biggest year-over-year increase since the summer of 2017.
Net income, excluding certain items, was $37 million, or 5 cents a share. That compares with $58 million, or 8 cents a share, a year earlier, Twitter said.
The YouTubers Union Is Not Messing Around
An organized labor movement for YouTubers is stirring in Europe, reports Edward Ongweso Jr.:
The YouTubers Union, a community-based movement fighting for the rights of content creators and users, has joined forces with IG Metall, Germany’s largest union and Europe’s largest trade union. Together, they have launched a joint venture called FairTube and sent a letter of demands to YouTube accompanied by a video explaining their concerns, demands, and plan of action.
The move is one of the most significant organized labor actions taken by creators on the platform, and puts some actual union power behind what has thus far been a nascent and disorganized movement.
Bollywood Rapper Sets Viewer Record YouTube Isn’t Talking About
Lucas Shaw explains how moneyed YouTubers are gaming platform metrics by buying ads:
When releasing a new single, major record labels will buy an advertisement on YouTube that places their music video in between other clips. If viewers watch the ad for more than few seconds, YouTube counts that as a view, boosting the overall total. Blackpink and Swift, among others, have done it. Badshah just took it a step further, people familiar with the matter say.
The practice creates doubts about the real popularity of these clips and reveals some of the murky ways in which artists and their labels promote their music—especially in emerging markets. YouTube, a subsidiary of Alphabet Inc.’s Google, is now reevaluating the way it judges records, according to two people familiar with the company’s thinking.
Now Even Funerals Are Livestreamed—and Families Are Grateful
Here’s a story from Paris Martineau about a phenomenon that seems inevitable, useful, and a little sad:
In a culture obsessed with tweeting and Instagramming every moment of life, it’s little surprise that streaming extends to death. Funeral livestreaming services have been around for more than a decade, but the practice has recently exploded in popularity, says Bryant Hightower, president-elect of the National Funeral Directors Association. He estimates that nearly 20 percent of US funeral homes now offer the service—a big number in an industry resistant to change—in response to demand from clients. Tech-savvy entrepreneurs offer livestreaming as a service to hesitant funeral directors.
Snapchat rolls out first major global ad campaign on back of surging user growth
Snap has a new ad campaign dedicated to celebrating “real friends,” as opposed to whoever follows you on Instagram. Cheekily, Snap paid influencers to post Snap ads on Instagram, getting their message out to a core audience without having to pay Facebook for the privilege.
Playlist
My friend Jason del Rey has a deeply reported new podcast out about the rise of Amazon. The second episode, out today, chronicles the development Alexa. Give it a listen!
Launches
Pinterest’s New Search Tool Puts Stress Relief in Your Feed
It’s been a slow week for launches to date, so here’s one from my vacation: a new anti-anxiety initiative from Pinterest built into the product itself:
When you type in an anxiety-related query—something like “work anxiety,” or “dealing with stress”—Pinterest will now display a box above the stream of pins. “If you’re feeling sad or stressed, here are some resources that may help improve your mood,” it says, above a disclaimer that notes Pinterest’s exercises are not a replacement for professional care. You can click into the box to see more, or scroll down to just look at the pins.
Takes
I enjoyed Matt Levine’s take on whether Facebook’s new privacy oversight committee will be independent in practice given Mark Zuckerberg’s controlling share of the stock:
Again I think Chopra is probably right that, practically speaking, there are serious limits on how aggressive and effective the committee can be in protecting user privacy. And as I suggested yesterday, if you think that Facebook and other big internet companies are systemically important institutions that should have formal responsibilities to the public as well as to shareholders, the way to do that is probably through carefully considered general legislation rather than through a negotiated settlement that binds only Facebook. But for the FTC to take some of Facebook’s governance out of shareholders’ (Zuckerberg’s) hands, even in a limited and symbolic way, suggests that the FTC does think that, and that Facebook concedes that it’s true. If you think that Facebook is too big to fail and needs to be regulated as a public utility, well, you won’t be happy with this settlement really, but it is a very teeny first step in that direction.
And finally ...
Do click through to check out the visual metaphor here, in which cotton candy dissolves immediately upon hitting the water. It is, as they say on Reddit, oddly satisfying:
this video perfectly describes how it feels to make stuff and put it online. https://t.co/HDdQHv6jkl
— Donwill® 'One Word No Space' Out Now (@donwill) July 30, 2019
Talk to me
Send me tips, comments, questions, and ideas for building brain-computer interfaces: casey@theverge.com.
Comments
Post a Comment