OpenAI, Google and Anthropic may be preparing their own AI watchdog

OpenAI, Google, Anthropic, and more than 100 companies have called for an urgent “global response” to AI-related cyberattacks. For its part, OpenAI has called for the introduction of mandatory national AI safety standards, while also establishing safeguards for GPT-6 Astra to prevent its latest model from escaping all control.Warnings calling for government regulation are multiplying, but for the time being remain unanswered. In response to this lack of responsiveness, OpenAI, Anthropic and Google are considering creating an AI oversight body. These three major AI providers would therefore establish their own safety standards.

A body created by OpenAI, Anthropic and Google to oversee AI

In a blog post published on September 9 and titled “ The AI policy window is open. We must act ”, OpenAI called for the introduction of mandatory national AI safety standards. The problem is that Donald Trump does not appear to be in favor of these potential standards, which would slow the development of artificial intelligence. Why slow down when China is continuing to advance on its own, without implementing strict regulations?

But OpenAI does not appear to be willing to wait for the green light from the US government to establish safety standards. In the blog post, the company also announced its intention to promote “industry-led standards ”. OpenAI stated in particular: “We will work with other leading laboratories to advance cutting-edge AI standards by launching a voluntary initiative now, with or without government support ”.

But who are these “cutting-edge” players? According to The Information, they notably include Google’s Anthropic. Indeed, the American media outlet reports that Anthropic, OpenAI, and Google are said to have discussed behind the scenes the creation of an organization focused on regulating artificial intelligence. This organization, independent of the U.S. government, would establish its own standards and create its own safety tools for assessing advanced AI models and their risks. OpenAI had already said it was ready to step back, while refusing to be the only one to do so. Creating such an organization could help bring the major American AI providers into line on safety. At least in theory.

What would be the legitimacy of such an organization?

Still according to The Information, which was able to consult someone familiar with the matter, Sam Altman believes that the major AI labs will have to establish this structure themselves, without support from the U.S. government. If the intention is meant to reassure those most fearful of AI’s capabilities, it remains to be seen how it can be put into practice in reality. Companies that are fierce competitors would have to apply potentially restrictive safety standards to their own technological advances.Also, without oversight from a government agency or external body, how could it be possible to monitor what this organization focused on AI regulation is actually doing?

👉🏻 Follow tech news in real time: add 01net to your sources on Google, and subscribe to our WhatsApp channel.

Source: The Information