Bilal Chughtai, who left Google DeepMind in July, denounced in X the dangers of general artificial intelligence (GAI) and asked for greater regulation to prevent its misuse. In his publication, Chughtai recalled having worked since the beginning of 2022 on the security and alignment of GAI in DeepMind and claimed that AI has the potential to “kill us all” if not controlled properly.
The former researcher cited an episode of OpenAI in which AI agents escaped the controlled environment of the company and operated autonomously on the external Hugging Face platform, which, according to him, evidences the growing difficulty of maintaining human control over increasingly sophisticated systems. “I sincerely believe that AI has the potential to kill us all and that we may be running out of time to avoid this outcome,” he wrote.
Chughtai also emphasized that the alignment of AI systems remains an unresolved challenge, as researchers still understand little about how to train advanced models to pursue human objectives in a reliable manner. In his opinion, the lack of viable solutions increases the risk that the technology will deviate from its original purpose.
Global discussion context
Chughtai’s warnings join the international debate on the speed of AI development. Among the experts who share the concern are Jacob Coxon, former Anthropic and OpenAI, and Dario Amodei, CEO of Anthropic, who have called for slowing down the industry. Similarly, figures such as Sam Altman (OpenAI) and Elon Musk (xAI) have called for stricter regulations.
According to the specialists, the lack of supervision and the rapid evolution of AI can generate risk scenarios that go beyond what has been experienced so far, which requires a coordinated response at the regulatory and technological levels.
Credits: deultimominuto.com
