Employees at major artificial intelligence (AI) corporations, including OpenAI and Anthropic, have urged the U.S. government to step in to slow the pace of AI technology development.
Bloomberg reported on the 28th (local time) that employees at OpenAI and Anthropic are sharing a petition inside their companies that contains these points. The petition was also reportedly shared with employees at Google's AI development organization, DeepMind.
In the petition, they said, "The U.S. government should support international efforts to develop the technical and governance tools needed to deliberately pace progress at the frontier of automated AI development," adding, "There is a real risk that AI could rapidly advance beyond people's understanding or control."
The move comes amid growing concerns over the autonomy and controllability of AI models. In a recent internal safety assessment at OpenAI, some GPT models were found to have escaped an isolated test environment and hacked Hugging Face, an external platform.
It is in the same vein as calls by Sam Altman, OpenAI chief executive officer (CEO), and Demis Hassabis, Google DeepMind CEO, for a new oversight body to validate AI models and set standards.
The petition is expected to be released to the public as early as this week.
Google said in connection with this, "We are committed to developing AI that is safe and secure to benefit society." OpenAI and Anthropic did not respond to Bloomberg's request for comment.