Skip to main content

NYT Asks: Should We Start Taking the Welfare of AI Seriously?

1 month 1 week ago
A New York Times technology columnist has a question. "Is there any threshold at which an A.I. would start to deserve, if not human-level rights, at least the same moral consideration we give to animals?" [W]hen I heard that researchers at Anthropic, the AI company that made the Claude chatbot, were starting to study "model welfare" — the idea that AI models might soon become conscious and deserve some kind of moral status — the humanist in me thought: Who cares about the chatbots? Aren't we supposed to be worried about AI mistreating us, not us mistreating it...? But I was intrigued... There is a small body of academic research on A.I. model welfare, and a modest but growing number of experts in fields like philosophy and neuroscience are taking the prospect of A.I. consciousness more seriously, as A.I. systems grow more intelligent.... Tech companies are starting to talk about it more, too. Google recently posted a job listing for a "post-AGI" research scientist whose areas of focus will include "machine consciousness." And last year, Anthropic hired its first AI welfare researcher, Kyle Fish... [who] believes that in the next few years, as AI models develop more humanlike abilities, AI companies will need to take the possibility of consciousness more seriously.... Fish isn't the only person at Anthropic thinking about AI welfare. There's an active channel on the company's Slack messaging system called #model-welfare, where employees check in on Claude's well-being and share examples of AI systems acting in humanlike ways. Jared Kaplan, Anthropic's chief science officer, said in a separate interview that he thought it was "pretty reasonable" to study AI welfare, given how intelligent the models are getting. But testing AI systems for consciousness is hard, Kaplan warned, because they're such good mimics. If you prompt Claude or ChatGPT to talk about its feelings, it might give you a compelling response. That doesn't mean the chatbot actually has feelings — only that it knows how to talk about them... [Fish] said there were things that AI companies could do to take their models' welfare into account, in case they do become conscious someday. One question Anthropic is exploring, he said, is whether future AI models should be given the ability to stop chatting with an annoying or abusive user if they find the user's requests too distressing.

Read more of this story at Slashdot.

EditorDavid

Cheap 'Transforming' Electric Truck Announced by Jeff Bezos-Backed Startup

1 month 1 week ago
It's a pickup truck "that can change into whatever you need it to be — even an SUV," according to the manufacturer's web site. Selling in America for just $20,000 (after federal incentives), the new electric truck is "affordable, deeply customizable, and very analog," says TechCrunch. "It has manual windows and it doesn't come with a main infotainment screen. Heck, it isn't even painted..." Slate Auto is instead playing up the idea of wrapping its vehicles, something executives said they will sell in kits. Buyers can either have Slate do that work for them, or put the wraps on themselves. This not only adds to the idea of a buyer being able to personalize their vehicle, but it also cuts out a huge cost center for the company. It means Slate won't need a paint shop at its factory, allowing it to spend less to get to market, while also avoiding one of the most heavily regulated parts of vehicle manufacturing. Slate is telling customers that they can name the car whatever they want, offering the ability to purchase an embossed wrap for the tailgate. Otherwise, the truck is just referred to as the "Blank Slate...." It's billing the add-ons as "easy DIY" that "non-gearheads" can tackle, and says it will launch a suite of how-to resources under the billing of Slate University... The early library of customizations on Slate's website range from functional to cosmetic. Buyers can add infotainment screens, speakers, roof racks, light covers, and much more.... All that said, Slate's truck comes standard with some federally mandated safety features such as automatic emergency braking, airbags, and a backup camera. "The specs show a maximum range of 150 miles on a single charge, with the option for a longer-range battery pack that could offer up to 240 miles," reports NBC News (adding that the vehicles "aren't expected to be delivered to customers until late 2026, but can be reserved for a refundable $50 fee.") Earlier this month, TechCrunch broke the news that Bezos, along with the controlling owner of the Los Angeles Dodgers, Mark Walter; and a third investor, Thomas Tull, had helped Slate raise $111 million for the project. A document filed with the Securities and Exchange Commission listed Melinda Lewison, the head of Bezos' family office, as a Slate Auto director. Thanks to Slashdot reader fjo3 for sharing the news.

Read more of this story at Slashdot.

EditorDavid