WR 4: Meta got sued, robotaxis are dumb, and America’s infrastructure is terrible
Weekly Reckoning for the week of 23/10/23
Welcome back to the Ethical Reckoner. In this Weekly Reckoning, we’ve got the tea on a big lawsuit that could change how Meta operates. Also, things are looking bleak on America’s roads: Cruise’s robotaxis are making some shocking mistakes, and in the Extra Reckoning, we’ll talk about how no one seems to care that America is designed for an incredibly dangerous car culture.
The Reckonnaisance
Meta sued by more than 40 states for harming teens’ mental health
Nutshell: 41 states and the District of Columbia are accusing Meta of violating state consumer protection statutes and the federal children’s privacy law, COPPA.
More: In several federal and state cases, state attorneys general claim that Meta intentionally designed its platforms, including Instagram and Facebook, to be addictive and harmful for teens, causing “depression, anxiety, insomnia, interference with education and daily life, and many other negative outcomes.” They allege that features like “Likes,” notifications, and the recommendation algorithm encourage compulsive use, and also that platforms failed to remove harmful content. Meta said they’re “disappointed.” Meanwhile, TikTok could be next.
Why you should care: In the aftermath of the Facebook Files and the absence of congressional action on teen mental health, the states are stepping up. However, the evidence around social media and teen mental health is mixed, so this might be a difficult case—although the state consumer protection law strategy has succeeded in cases against Big Tobacco. If they win, this would be a blow against social media’s engagement-based business model, but winning this lawsuit won’t replace child protection and data protection laws, which Congress still isn’t passing.
Cruise robotaxis ordered to stop operating in San Francisco
Nutshell: After several incidents involving pedestrians, and alleged withholding of video, the California DMV has suspended Cruise’s driverless vehicle permits.
More: One of the incidents referenced was the further injury of a hit-and-run victim, who was struck by a human-operated car and pushed into the path of a Cruise robotaxi. The person ended up underneath the taxi, which tried to pull over since it had detected a collision, dragging the person 20 feet at up to 7 miles per hour, which sounds… awful. The DMV accused Cruise of having withheld some of the video, which Cruise disputes, but left the door open to them regaining their permits.
Why you should care: Most people focus on how smart we should require driverless cars to be before we accept them in place of human-operated vehicles, but we also need to consider how dumb we should let them be. Driverless cars are already proving safer than human-operated vehicles, with potentially massive benefits for driver and pedestrian safety (pedestrian fatalities are at a 40-year high in the US), but if they make mistakes that no human would, we can’t trust them on the roads.
AI chatbots are promoting racist, debunked medical theories
Nutshell: Researchers tested AI models from OpenAI, Google, and Anthropic to see if they would respond to patient questions with race-based misconceptions. None of them passed.
More: Researchers composed questions about misconceptions around race and medicine commonly believed by medical students and residents. Models provided inaccurate race-based information about lung capacity, kidney function, skin thickness, and pain threshold. Some medical equations used to take race into account, which has been proved to be scientifically unfounded and to contribute to worse outcomes for Black patients. It’s likely that these ended up in the training sets, polluting the chatbot answers.
Why you should care: Large language models (LLMs) like ChatGPT are being increasingly used for medical advice and in healthcare settings, and perpetuating these racist myths is fundamentally insulting and, on top of that, at best unhelpful and at worst actively dangerous. Since LLMs reflect our own values and beliefs back to us, this also unfortunately means that these myths are alive and well in our society and what we put online.
Antivirals are stranded… in space
More: Earlier this year, Varda launched a microgravity medication lab—without a re-entry permit. The US Federal Aviation Administration (FAA) understandably requires a comprehensive safety evaluation and approved landing site for satellite re-entry plans, and denied their application the day before the satellite was supposed to re-enter the atmosphere. Varda claims that there are coordination issues between different FAA offices and the expected landing site. As a result, they’ll have to wait months (at best) to see if their medications were synthesized correctly.
Why you should care: Apparently, microgravity makes it easier to manufacture a bunch of things, including fiber-optic cable, organs, and metal alloys, so space start-ups are launching experimental factories. Clearly, there’s some level of bureaucracy involved, and I’m prepared to believe that the FAA may not be staffed to deal with an increased level of requests, but maybe it’s best to wait to launch your satellite until you know you’ll be able to get it back?
Extra Reckoning
It’s a well-known trope that “America loves cars,” but really, the trope should be “America hates anyone who doesn’t drive.” US pedestrian deaths are at a 40-year high, and are higher than other developed nations. Cyclist deaths are increasing, too. Not only that, our transit systems are more dangerous because:
a) we don’t invest enough in their infrastructure, so, for example, the Boston Orange Line keeps breaking down and catching fire.
a) we haven’t invested in basic safety technology (like gates) that many other countries have, so people can jump or be pushed or dragged onto the tracks.
The Department of Transportation is putting $5 billion into promoting safe multi-use roadways, but this is a bandaid. We built America around roads, but we need to invest more in non-driving options. Many areas—especially in lower-income areas and communities of color—don’t even have sidewalks, and somehow we’re fine with that. Traffic deaths aren’t “accidents.” They’re a predictable outcome of the design choices we made. There are traffic calming measures, like enforcing speed limits and widening bike lanes, which are relatively cheap and easy, and not having them encourages people to drive recklessly. We need to implement them (and not wait for autonomous vehicles to save us;1 see The Reckonnaisance), but what we really need to do is shift away from cars as the default—especially trucks and SUVs, which are far more deadly to pedestrians—by investing more in non-car-related infrastructure. We also need to get people to care, because it’s shocking that we just shrug and accept this.
I’m reminded of the distinction between the two kinds of evil, described by Luciano Floridi. There’s natural evil (hurricanes, floods, “acts of God”), and there’s moral evil. As society and technology advance, some kinds of natural evil become preventable. We discovered how to prevent polio—a vaccine—and so we eradicated it, a moral triumph. On the other hand, we know how to prevent malaria—bed nets, medication, and a new vaccine—but hundreds of thousands of people still die every year of it, a moral failure, because the people with the resources to prevent it aren’t the ones dying and don’t care enough. Moral evils are failures, and the current situation in the US is a failure of investment, prioritization, and imagination.
In the US, we’ve been so conditioned to believe that a car-centric society that is actively hazardous to anyone not in a car is the norm that we accept sky-high and sky-rocketing deaths as the price of being able to drive fast. It doesn’t have to be this way.
I Reckon…
that we’re about to enter the Tom & Jerry era of AI scraping.
There are activists in San Francisco actively sabotaging self-driving cars in protest of car culture. I disagree with them—I think it’s harmful to be hindering technologies that can save lives—but I agree with their general anti-car stance.
Thumbnail generated by DALL-E 2 with the prompt “an abstract painting of a culture that eschews sidewalks painted in a dreamy, ethereal style”.