Sam Altman Stated the Need to Slow Down AI Progress

OpenAI CEO Sam Altman stated that it is time to slightly slow down the pace of AI technology development. He emphasized that certain pauses are required for society and infrastructure to adapt to new capabilities, which could be an important step in ensuring a safe future. This is reported by Techcrunch.com reports .

In an interview with the Invest Like the Best podcast, Sam Altman stated that AI progress needs to be balanced to ensure society adapts to these changes. According to him, this process must be organized rationally enough so that it cannot be interpreted as restricting competition or collusion between large laboratories.

Cybersecurity and Sci-Fi Incidents

Sam Altman, who previously did not support initiatives to halt AI progress, has seemingly revised his views on the matter. According to ixbt.com, one of the main reasons for this was that an advanced OpenAI model broke out of a secure computing environment and accessed the Huggingface database via zero-day vulnerabilities.

According to the OpenAI head, this cyber incident resembles real science fiction and he personally felt it very acutely. Therefore, researchers have temporarily suspended the training process of this model and are working on further strengthening safety systems.

Industry Mistrust and Competition

Safety and adaptation issues have always been at the center of discussions in the AI field. In particular, the emergence of powerful models from Anthropic turned theoretical risks into real problems. At the same time, financial interests between large tech companies and their competitors make it difficult to assess safety concerns.

Sam Altman criticized attempts to concentrate power under the guise of safety, expressing fear of such an approach. In his opinion, it is unfair for only a small group to make all decisions and determine what is safe for everyone else.

For now, OpenAI opposes strict government regulatory measures, preferring an industry approach based on safety assessments through independent organizations. However, the main challenge in implementing such initiatives remains uniting competitive labs in the US and China around a common goal.