There are some people who are responsible for studying AI and deciding in what ways it can, essentially, destroy the world. These people work at Anthropic as part of a group called the social impact team, which Hayden spent time with for a profile she published this week.
See also: Anthropic releases Claude Opus 4.5 with improved features
The team consists of just nine people out of the more than 2,000 who work at Anthropic. Their sole job, as the team members themselves say, is to investigate and publish “inconvenient truths” about how people use AI tools, what chatbots might be doing to our mental health, and how all of this might have broader implications for the job market, the economy, and even our elections.
This, of course, raises a number of problems. The most important is whether this group can remain independent, or even exist at all, as it publishes findings about Anthropic’s products that may be unpleasant or politically charged. Moreover, there is a lot of pressure on the AI industry in general and Anthropic in particular to comply with the Trump administration, which issued an executive order in July banning so-called “woke AI.”
See also: ServiceNow AI Agents: Abuse of default settings for prompt injection
If you follow the tech industry, the outline of this story will sound familiar. We’ve seen this most recently with social media companies and the trust and safety groups responsible for curating content. Meta went through countless cycles of this, where it devoted resources to solving problems created by its own scale and the unpredictable nature of products like Facebook and Instagram. And then, after a while, it seems like the resources ran out, or Mark Zuckerberg got bored or more interested in MMA or just wanted to get closer to Trump, and the products didn’t really change to reflect what the research showed.
See also: Vulnerability in Claude AI exposes enterprise data
We’re living in one of those moments right now. Social platforms have cut back on investments in election integrity and other forms of content management. Meanwhile, Silicon Valley is working closely with the Trump White House to resist meaningful efforts to regulate AI.
