Report from the Washington Post
In Brief – OpenAI and Anthropic have endorsed a petition urging the US government to help create an international regime to slow down and regulate the pace of “automated AI development.” It is signed by over 1,200 “employees of frontier AI companies” including Anthropic, OpenAI, Google, Microsoft, NVIDIA and Meta. The group is focused on the expected point when AI systems can automate AI research itself, a developmental priority of both Anthropic and OpenAI that some believe may come soon, which the signatories say could overtake humanity’s ability to safely control the technology. While they note that companies and countries are “under intense competitive pressure not to unilaterally slow” AI development, they call on the US Government to work internationally on technical and governance tools to regulate development.
Context – What is the point of the expert petitions calling on government to ward off AI doom? In early 2023, just a couple of months after the public release of ChatGPT, thousands of very smart people, including tech world notables Elon Musk and Steve Wozniak, called on governments to quickly impose a six-month moratorium on AI development because of the risks. That was never going to happen and it didn’t. Just a couple of weeks ago a group of experts led by economists, including 15 Nobel laureates, released a petition calling on policymakers and technology leaders to work to understand the economics of AI and “build the incentives, guardrails, and institutions needed to steer AI” to complement humans and benefits society. At least the economist-led group mentioned a 10-year time horizon for the possible calamities, which is much safer given how near-term predictions about AI capabilities and impacts keep missing. Yet, AI doomerism is so prevalent. Writer and analyst Derek Thompson has said that even the most knowledgeable AI experts quickly veer into science fiction narratives because nobody knows how the tech will evolve, let alone the world around it. It’s not surprising because everyone involved has been raised in a culture of mostly dystopian AI science fiction. Maybe this also explains where they learned about how government works.