Four of the biggest names in artificial intelligence—Anthropic, OpenAI, Google and Microsoft—are reportedly discussing a plan to police themselves before regulators do it for them. The idea centers on a shared code of conduct and an independent safety body that would monitor how quickly these companies push the boundaries of AI capability.
The move comes amid growing unease over “superintelligence,” the term researchers use for AI systems that could eventually surpass human reasoning across nearly every domain. Critics have warned that fierce competition between labs to release ever more powerful models leaves little room for caution, testing, or reflection on potential risks.
By proposing common safety standards, the companies appear to be trying to slow that competitive pressure without waiting for governments to impose binding rules. A dedicated oversight body could, in theory, evaluate new systems, flag dangerous capabilities, and give the public more visibility into how these models are built and deployed.
It’s a notable gesture from firms that have often resisted heavy-handed regulation, yet it also raises questions about whether an industry-led pact can carry real teeth, or whether it mainly serves to preempt tougher government oversight.
To read the full report and see how this proposal might take shape, check out the original story at jornada.com.mx.
Source: Anthropic, OpenAI, Google y Microsoft buscan autoregular su IA: plantean código de conducta y organismo de seguridad (jornada.com.mx). English version produced with AI assistance.
Image: Hkhatun6564.