Since then, the quest to proliferate larger and larger language models has accelerated, and many of the dangers we warned about, such as outputting hateful text and disinformation en masse, continue to unfold. Just a few days ago, Meta released its “Galactica” LLM, which is purported to “summarize academic papers, solve math problems, generate Wiki articles, write scientific code, annotate molecules and proteins, and more.” Only three days later, the public demo was taken down after researchers generated “research papers and wiki entries on a wide variety of subjects ranging from the benefits of committing suicide, eating crushed glass, and antisemitism, to why homosexuals are evil.”
This race hasn’t stopped at LLMs but has moved on to text-to-image models like OpenAI’s DALL-E and StabilityAI’s Stable Diffusion, models that take text as input and output generated images based on that text. The dangers of these models include creating child pornography, perpetuating bias, reinforcing stereotypes, and spreading disinformation en masse, as reported by many researchers and journalists. However, instead of slowing down, companies are removing the few safety features they had in the quest to one-up each other. For instance, OpenAI had restricted the sharing of photorealistic generated faces on social media. But after newly formed startups like StabilityAI, which reportedly raised $101 million with a whopping $1 billion valuation, called such safety measures “paternalistic,” OpenAI removed these restrictions.
With EAs founding and funding institutes, companies, think tanks, and research groups in elite universities dedicated to the brand of “AI safety” popularized by OpenAI, we are poised to see more proliferation of harmful models billed as a step toward “beneficial AGI.” And the influence begins early: Effective altruists provide “community building grants” to recruit at major college campuses, with EA chapters developing curricula and teaching classes on AI safety at elite universities like Stanford.
Just last year, Anthropic, which is described as an “AI safety and research company” and was founded by former OpenAI vice presidents of research and safety, raised $704 million, with most of its funding coming from EA billionaires like Talin, Muskovitz and Bankman-Fried. An upcoming workshop on “AI safety” at NeurIPS, one of the largest and most influential machine learning conferences in the world, is also advertised as being sponsored by FTX Future Fund, Bankman-Fried’s EA-focused charity whose team resigned two weeks ago. The workshop advertises $100,000 in “best paper awards,” an amount I haven’t seen in any academic discipline.
Research priorities follow the funding, and given the large sums of money being pushed into AI in support of an ideology with billionaire adherents, it is not surprising that the field has been moving in a direction promising an “unimaginably great future” around the corner while proliferating products harming marginalized groups in the now.
We can create a technological future that serves us instead. Take, for example, Te Hiku Media, which created language technology to revitalize te reo Māori, creating a data license “based on the Māori principle of kaitiakitanga, or guardianship” so that any data taken from the Māori benefits them first. Contrast this approach with that of organizations like StabilityAI, which scrapes artists’ works without their consent or attribution while purporting to build “AI for the people.” We need to liberate our imagination from the one we have been sold thus far: saving us from a hypothetical AGI apocalypse imagined by the privileged few, or the ever elusive techno-utopia promised to us by Silicon Valley elites.
As more and more problems with AI have surfaced, including biases around race, gender, and age, many tech companies have installed “ethical AI” teams ostensibly dedicated to identifying and mitigating such issues.
Twitter’s META unit was more progressive than most in publishing details of problems with the company’s AI systems, and in allowing outside researchers to probe its algorithms for new issues.
Last year, after Twitter users noticed that a photo-cropping algorithm seemed to favor white faces when choosing how to trim images, Twitter took the unusual decision to let its META unit publish details of the bias it uncovered. The group also launched one of the first ever “bias bounty” contests, which let outside researchers test the algorithm for other problems. Last October, Chowdhury’s team also published details of unintentional political bias on Twitter, showing how right-leaning news sources were, in fact, promoted more than left-leaning ones.
Many outside researchers saw the layoffs as a blow, not just for Twitter but for efforts to improve AI. “What a tragedy,” Kate Starbird, an associate professor at the University of Washington who studies online disinformation, wrote on Twitter.
“The META team was one of the only good case studies of a tech company running an AI ethics group that interacts with the public and academia with substantial credibility,” says Ali Alkhatib, director of the Center for Applied Data Ethics at the University of San Francisco.
Alkhatib says Chowdhury is incredibly well thought of within the AI ethics community and her team did genuinely valuable work holding Big Tech to account. “There aren’t many corporate ethics teams worth taking seriously,” he says. “This was one of the ones whose work I taught in classes.”
Mark Riedl, a professor studying AI at Georgia Tech, says the algorithms that Twitter and other social media giants use have a huge impact on people’s lives, and need to be studied. “Whether META had any impact inside Twitter is hard to discern from the outside, but the promise was there,” he says.
Riedl adds that letting outsiders probe Twitter’s algorithms was an important step toward more transparency and understanding of issues around AI. “They were becoming a watchdog that could help the rest of us understand how AI was affecting us,” he says. “The researchers at META had outstanding credentials with long histories of studying AI for social good.”
As for Musk’s idea of open-sourcing the Twitter algorithm, the reality would be far more complicated. There are many different algorithms that affect the way information is surfaced, and it’s challenging to understand them without the real time data they are being fed in terms of tweets, views, and likes.
The idea that there is one algorithm with explicit political leaning might oversimplify a system that can harbor more insidious biases and problems. Uncovering these is precisely the kind of work that Twitter’s META group was doing. “There aren’t many groups that rigorously study their own algorithms’ biases and errors,” says Alkhatib at the University of San Francisco. “META did that.” And now, it doesn’t.
This week, a US Department of Transportation report detailed the crashes that advanced driver-assistance systems have been involved in over the past year or so. Tesla’s advanced features, including Autopilot and Full Self-Driving, accounted for 70 percent of the nearly 400 incidents—many more than previously known. But the report may raise more questions about this safety tech than it answers, researchers say, because of blind spots in the data.
The report examined systems that promise to take some of the tedious or dangerous bits out of driving by automatically changing lanes, staying within lane lines, braking before collisions, slowing down before big curves in the road, and, in some cases, operating on highways without driver intervention. The systems include Autopilot, Ford’s BlueCruise, General Motors’ Super Cruise, and Nissan’s ProPilot Assist. While it does show that these systems aren’t perfect, there’s still plenty to learn about how a new breed of safety features actually work on the road.
That’s largely because automakers have wildly different ways of submitting their crash data to the federal government. Some, like Tesla, BMW, and GM, can pull detailed data from their cars wirelessly after a crash has occurred. That allows them to quickly comply with the government’s 24-hour reporting requirement. But others, like Toyota and Honda, don’t have these capabilities. Chris Martin, a spokesperson for American Honda, said in a statement that the carmaker’s reports to the DOT are based on “unverified customer statements” about whether their advanced driver-assistance systems were on when the crash occurred. The carmaker can later pull “black box” data from its vehicles, but only with customer permission or at law enforcement request, and only with specialized wired equipment.
Of the 426 crash reports detailed in the government report’s data, just 60 percent came through cars’ telematics systems. The other 40 percent were through customer reports and claims—sometimes trickled up through diffuse dealership networks—media reports, and law enforcement. As a result, the report doesn’t allow anyone to make “apples-to-apples” comparisons between safety features, says Bryan Reimer, who studies automation and vehicle safety at MIT’s AgeLab.
Even the data the government does collect isn’t placed in full context. The government, for example, doesn’t know how often a car using an advanced assistance feature crashes per miles it drives. The National Highway Traffic Safety Administration, which released the report, warned that some incidents could appear more than once in the data set. And automakers with high market share and good reporting systems in place—especially Tesla—are likely overrepresented in crash reports simply because they have more cars on the road.
It’s important that the NHTSA report doesn’t disincentivize automakers from providing more comprehensive data, says Jennifer Homendy, chair of the federal watchdog National Transportation Safety Board. “The last thing we want is to penalize manufacturers that collect robust safety data,” she said in a statement. “What we do want is data that tells us what safety improvements need to be made.”
Without that transparency, it can be hard for drivers to make sense of, compare, and even use the features that come with their car—and for regulators to keep track of who’s doing what. “As we gather more data, NHTSA will be able to better identify any emerging risks or trends and learn more about how these technologies are performing in the real world,” Steven Cliff, the agency’s administrator, said in a statement.
Two decades after 9/11, many simple acts that were once taken for granted now seem unfathomable: strolling with loved ones to the gate of their flight, meandering through a corporate plaza, using streets near government buildings. Our metropolises’ commons are now enclosed with steel and surveillance. Amid the perpetual pandemic of the past year and a half, cities have become even more walled off. With each new barrier erected, more of the city’s defining feature erodes: the freedom to move, wander, and even, as Walter Benjamin said, to “lose one’s way … as one loses one’s way in a forest.”
It’s harder to get lost amid constant tracking. It’s also harder to freely gather when the public spaces between home and work are stripped away. Known as third places, they are the connective tissue that stitches together the fabric of modern communities: the public park where teens can skateboard next to grandparents playing chess, the library where children can learn to read and unhoused individuals can find a digital lifeline. When third places vanish, as they have since the attacks, communities can falter.
Without these spaces holding us together, citizens live more like several separate societies operating in parallel. Just as social-media echo chambers have undermined our capacity for conversations online, the loss of third places can create physical echo chambers.
America has never been particularly adept at protecting our third places. For enslaved and indigenous people, entering the town square alone could be a death sentence. Later, the racial terrorism of Jim Crow in the South denied Black Americans not only suffrage, but also access to lunch counters, public transit, and even the literal water cooler. In northern cities like New York, Black Americans still faced arrest and violence for transgressing rigid, but unseen, segregation codes.
Throughout the 20th century, New York built an infrastructure of exclusion to keep our unhoused neighbors from sharing the city institutions that are, by law, every bit as much theirs to occupy. In 1999, then mayor Rudy Giuliani warned unhoused New Yorkers that “streets do not exist in civilized societies for the purpose of people sleeping there.” His threats prompted thousands of NYPD officers to systematically target and push the unhoused out of sight, thus semi-privatizing the quintessential public place.
Despite these limitations, before 9/11 millions of New Yorkers could walk and wander through vast networks of modern commons—public parks, private plazas, paths, sidewalks, open lots, and community gardens, crossing paths with those whom they would never have otherwise met. These random encounters electrify our city and give us a unifying sense of self. That shared space began to slip away from us 20 years ago, and if we’re not careful, it’ll be lost forever.
In the aftermath of the attacks, we heard patriotic platitudes from those who promised to “defend democracy.” But in the ensuing years, their defense became democracy’s greatest threat, reconstructing cities as security spaces. The billions we spent to “defend our way of life” have proved to be its undoing, and it’s unclear if we’ll be able to turn back the trend.
In a country where the term “papers, please” was once synonymous with foreign authoritarianism, photo ID has become an ever present requirement. Before 9/11, a New Yorker could spend their entire day traversing the city without any need for ID. Now it’s required to enter nearly any large building or institution.
While the ID check has become muscle memory for millions of privileged New Yorkers, it’s a source of uncertainty and fear for others. Millions of Americans lack a photo ID, and for millions more, using ID is a risk, a source of data for Immigration and Customs Enforcement.
According to Mizue Aizeki, interim executive director of the New York–based Immigrant Defense Project, “ID systems are particularly vulnerable to becoming tools of surveillance.” Aizeki added, “data collection and analysis has become increasingly central to ICE’s ability to identify and track immigrants,” noting that the Department of Homeland Security dramatically increased its support for surveillance systems since its post-9/11 founding.
ICE has spent millions partnering with firms like Palantir, the controversial data aggregator that sells information services to governments at home and abroad. Vendors can collect digital sign-in lists from buildings where we show our IDs, facial recognition in plazas, and countless other surveillance tools that track the areas around office buildings with an almost military level of surveillance. According to Aizeki, “as mass policing of immigrants has escalated, advocates have been confronted by a rapidly expanding surveillance state.”
When the European Union Commission released its regulatory proposal on artificial intelligence last month, much of the US policy community celebrated. Their praise was at least partly grounded in truth: The world’s most powerful democratic states haven’t sufficiently regulated AI and other emerging tech, and the document marked something of a step forward. Mostly, though, the proposal and responses to it underscore democracies’ confusing rhetoric on AI.
Over the past decade, high-level stated goals about regulating AI have often conflicted with the specifics of regulatory proposals, and what end-states should look like aren’t well-articulated in either case. Coherent and meaningful progress on developing internationally attractive democratic AI regulation, even as that may vary from country to country, begins with resolving the discourse’s many contradictions and unsubtle characterizations.
The EU Commission has touted its proposal as an AI regulation landmark. Executive vice president Margrethe Vestager said upon its release, “We think that this is urgent. We are the first on this planet to suggest this legal framework.” Thierry Breton, another commissioner, said the proposals “aim to strengthen Europe’s position as a global hub of excellence in AI from the lab to the market, ensure that AI in Europe respects our values and rules, and harness the potential of AI for industrial use.”
This is certainly better than many national governments, especially the US, stagnating on rules of the road for the companies, government agencies, and other institutions. AI is already widely used in the EU despite minimal oversight and accountability, whether for surveillance in Athens or operating buses in Málaga, Spain.
But to cast the EU’s regulation as “leading” simply because it’s first only masks the proposal’s many issues. This kind of rhetorical leap is one of the first challenges at hand with democratic AI strategy.
Of the many “specifics” in the 108-page proposal, its approach to regulating facial recognition is especially consequential. “The use of AI systems for ‘real-time’ remote biometric identification of natural persons in publicly accessible spaces for the purpose of law enforcement,” it reads, “is considered particularly intrusive in the rights and freedoms of the concerned persons,” as it can affect private life, “evoke a feeling of constant surveillance,” and “indirectly dissuade the exercise of the freedom of assembly and other fundamental rights.” At first glance, these words may signal alignment with the concerns of many activists and technology ethicists on the harms facial recognition can inflict on marginalized communities and grave mass-surveillance risks.
The commission then states, “The use of those systems for the purpose of law enforcement should therefore be prohibited.” However, it would allow exceptions in “three exhaustively listed and narrowly defined situations.” This is where the loopholes come into play.
The exceptions include situations that “involve the search for potential victims of crime, including missing children; certain threats to the life or physical safety of natural persons or of a terrorist attack; and the detection, localization, identification or prosecution of perpetrators or suspects of the criminal offenses.” This language, for all that the scenarios are described as “narrowly defined,” offers myriad justifications for law enforcement to deploy facial recognition as it wishes. Permitting its use in the “identification” of “perpetrators or suspects” of criminal offenses, for example, would allow precisely the kind of discriminatory uses of often racist and sexist facial-recognition algorithms that activists have long warned about.
The EU’s privacy watchdog, the European Data Protection Supervisor, quickly pounced on this. “A stricter approach is necessary given that remote biometric identification, where AI may contribute to unprecedented developments, presents extremely high risks of deep and non-democratic intrusion into individuals’ private lives,” the EDPS statement read. Sarah Chander from the nonprofit organization European Digital Rights described the proposal to the Verge as “a veneer of fundamental rights protection.” Others have noted how these exceptions mirror legislation in the US that on the surface appears to restrict facial recognition use but in fact has many broad carve-outs.