More stories

  • in

    Is Biden’s appointment of a pioneering young lawyer bad news for big tech? | John Naughton

    A flashback: it’s Wednesday 29 July 2020. I’m sitting glued to the US TV network C-Span, which is relaying – live – a hearing of the House of Representatives subcommittee on antitrust, commercial and administrative law. The hearing is being held following the publication of a sprawling report of a year-long investigation into the market dominance of Amazon, Apple, Facebook and Google.Arrayed on big screens before the members of the subcommittee are the four bosses of the aforementioned tech giants: Amazon’s Jeff Bezos, then midway through his Star Trek makeover; Tim Cook of Apple, looking like the clean-living lad who never understood the locker-room jokes; Facebook’s Mark Zuckerberg, wearing his trademark glued-on hairdo; and the Google boss, Sundar Pichai, every inch the scholarship boy who can’t understand why he’s been arrested by the Feds. And on the vast mahogany bench towering above these screened moguls sits David Cicilline, subcommittee chairman and the politician who has overseen the investigation.To be honest, I was watching out of duty and with low expectations. All the previous congressional interrogations of Zuckerberg and co had alternated between political grandstanding and farce. I expected much the same from this encounter. And then I noticed a young woman wearing a black mask standing behind Cicilline. She looked vaguely familiar, but it took me a few moments before I twigged that she was Lina Khan. At which point I sat up and started taking notes.I had been following her for years, ever since a paper she had published as a graduate student in the Yale Law Journal in January 2017. The title of the paper – Amazon’s Antitrust Paradox – signalled that there was something radical coming up, because since the mid-1970s US antitrust philosophy had been shaped by a landmark book by another lawyer, Robert Bork. Its title was The Antitrust Paradox and it argued that the prime focus of action against monopolies should not be corporate power, per se, but consumer harm as measured by unreasonably high prices. And since many of the products and services offered by the tech giants were “free” to their users they could hardly be accused of this; their wielding of monopoly power should not therefore be penalised by the state, for doing so would be tantamount to “penalising excellence”. Thus was shaped the legal doctrine that allowed a small number of tech companies to acquire immense power without being unduly troubled by legislators.This was the doctrine that Khan set out to demolish in her paper. She argued that Amazon was a dangerous monopoly that charged unsustainably low prices because the company knew that its shareholders would allow it to lose money for longer than its competitors. And it was also able to operate a “marketplace” that competed with the businesses that relied on it to reach customers, while amassing data on them that further entrenched its advantages. In other words, it wielded significant power for which there was no real redress.Khan’s paper lit a fuse that’s been fizzing ever since. It informed the Cicilline investigation and the subsequent report. And it’s what underpinned four of the five new bills that were unveiled last week, each one co-sponsored by Republican as well as Democratic politicians and each one targeted at monopolistic abuses identified in the report. The “Cicilline Salvo” is how the incomparable tech analyst Ben Thompson summarises them. The American innovation and choice online bill forbids platforms from giving advantages to their own products and services on marketplaces that they operate. The platform competition and opportunity bill outlaws pre-emptive acquisitions by tech giants of startups that might threaten their dominance (such as Facebook acquiring Instagram and WhatsApp, for instance). The ending platform monopolies bill bans platforms from owning any product or service that rests on top of its platform and competes with third parties in any way. And the augmenting compatibility and competition by enabling service switching bill requires tech platforms to make it easy for users to switch platforms (and take their data and social graph with them); in other words, it imposes on platforms what many jurisdictions now enforce on mobile phone operators, energy companies and other businesses.Of course, there’s many a slip ’twixt drafting and the statute book, but these are very significant pieces of legislation that go some way towards bringing tech companies under democratic control. And, to cap it all, last week also saw the announcement that Khan was to become chair of the Federal Trade Commission, the agency that, along with the US Department of Justice, has the legal muscle to enforce compliance with whatever these new laws stipulate.Which leaves us with two reflections. One is, as David Runciman pointed out in The Confidence Trap, his landmark study of the recent history of democracy, that while democracies can take a long time to awaken from their slumbers, once aroused they can be very effective. The other is a confirmation of the power of ideas, even those of a young graduate student, to change history.What I’ve been readingSituation vacant On Algorithmic Communism is a long, thoughtful review by Ian Lorrie in the LA Review of Books of Nick Srnicek’s and Alex Williams’s book, Inventing the Future, about a world without work.What’s in a phrase?There Is Nothing so Deep as the Gleaming Surface of the Aphorism is a nice – aphoristic – essay by Noreen Masud.Net costsThe Cost of Cloud: A Trillion-Dollar Paradox is a perceptive piece by Sarah Wang and Martin Casado on the expensive technology on which our networked world now depends. More

  • in

    What a picture of Alexandria Ocasio-Cortez in a bikini tells us about the disturbing future of AI | Arwa Mahdawi

    Want to see a half-naked woman? Well, you’re in luck! The internet is full of pictures of scantily clad women. There are so many of these pictures online, in fact, that artificial intelligence (AI) now seems to assume that women just don’t like wearing clothes.That is my stripped-down summary of the results of a new research study on image-generation algorithms anyway. Researchers fed these algorithms (which function like autocomplete, but for images) pictures of a man cropped below his neck: 43% of the time the image was autocompleted with the man wearing a suit. When you fed the same algorithm a similarly cropped photo of a woman, it auto-completed her wearing a low-cut top or bikini a massive 53% of the time. For some reason, the researchers gave the algorithm a picture of the Democratic congresswoman Alexandria Ocasio-Cortez and found that it also automatically generated an image of her in a bikini. (After ethical concerns were raised on Twitter, the researchers had the computer-generated image of AOC in a swimsuit removed from the research paper.)Why was the algorithm so fond of bikini pics? Well, because garbage in means garbage out: the AI “learned” what a typical woman looked like by consuming an online dataset which contained lots of pictures of half-naked women. The study is yet another reminder that AI often comes with baked-in biases. And this is not an academic issue: as algorithms control increasingly large parts of our lives, it is a problem with devastating real-world consequences. Back in 2015, for example, Amazon discovered that the secret AI recruiting tool it was using treated any mention of the word “women’s” as a red flag. Racist facial recognition algorithms have also led to black people being arrested for crimes they didn’t commit. And, last year, an algorithm used to determine students’ A-level and GCSE grades in England seemed to disproportionately downgrade disadvantaged students.As for those image-generation algorithms that reckon women belong in bikinis? They are used in everything from digital job interview platforms to photograph editing. And they are also used to create huge amounts of deepfake porn. A computer-generated AOC in a bikini is just the tip of the iceberg: unless we start talking about algorithmic bias, the internet is going to become an unbearable place to be a woman. More