OpenAI, Anthropic, and Google Explore Creating a Joint Body for AI Safety Standards
Discussions on establishing a joint body for AI safety standards
Translated by Qiyas from السياسة — Original source
OpenAI, Anthropic, and Google are exploring the creation of a joint body led by the artificial intelligence sector to establish safety standards, a move reflecting growing concerns within the industry about the rapid advancement of capabilities in advanced models and the difficulty of keeping protective measures pace with this development.
According to The Information, citing people familiar with the discussions, the three companies have held talks about collaborating to establish an AI standards body, and meetings of the project’s working group are ongoing.
Review demographic data
Subscribe to digital newspapers
The site reported that these discussions took place even before Anthropic CEO Dario Amodei called on AI companies to coordinate on testing the technology, auditing it, and establishing shared safety standards.
Reuters, citing Bloomberg, reported that OpenAI is working with Anthropic and Google DeepMind to coordinate their efforts on AI safety, amid increasing calls for implementing shared controls on the development of the most advanced models.
On Saturday, Amodei called on AI companies to slow the pace of developing advanced model capabilities, arguing that this would give risk mitigation measures a chance to keep up with the rapid technological evolution.
Amodei proposed engaging independent parties to evaluate systems, enhancing coordination among AI companies to set common safety standards, and fostering broader international cooperation to address potential risks.
OpenAI CEO Sam Altman quickly endorsed Amodei’s position, emphasizing the need to regulate the pace of advanced model development. Elon Musk, owner of xAI, also supported the call to slow down development.
In the context of ongoing discussions about the new body, The Information reported that Altman expressed support for creating an entity to test and audit AI systems, noting that major industry players may need to act together to establish a standards body.
