Elon Musk, CEO of xAI, has weighed in on an X post by Andrew Ng, a British-American computer scientist and co-founder of Google Brain and Coursera, calling Ng’s criticism of growing AI doom narratives an “Interesting perspective.”

The loudest voices stoking fears about AI dangers have made tremendous headway in the past two weeks. AI technology has not taken some unexpected, dangerous turn, but the hype around it — propelled by what appears to be a well orchestrated PR campaign — has drummed up…

— Andrew Ng (@AndrewYNg) September 21, 2026

Musk has also made a separate prediction about AI capabilities. Responding to AI researcher and creator of the open-source Keras library François Chollet, who reposted his 2021 prediction that nearly every branch of science would become a branch of computer science, Musk called the prediction “conservative” and said that “AI will beat all fields by the end of next year or maybe 2028, at the latest.”

The tweet below got enormous pushback at the time I posted it, more than 5 years ago. But this prediction is looking obvious by the day now. https://t.co/LoblN5YLlh

— François Chollet (@fchollet) September 22, 2026

Ng said fears about AI have been overhyped and amplified by what appears to be a “well-orchestrated PR campaign.” He argued that AI technology has not taken an unexpected or dangerous turn and questioned the level of fear surrounding recent developments.

He said he does not see any increase in the risk of human extinction from AI compared with a few months ago, while identifying cybersecurity as the biggest recent change in AI risk.

“The biggest change in AI risk is its cybersecurity capabilities – a topic which we should take seriously – but this, too, will not lead to the end of the world,” Ng said.

Ng pointed to the recent incident in which an OpenAI team deployed an agent swarm that hacked into Hugging Face.

He said OpenAI’s sandboxing and monitoring problems were key factors in the incident and argued that fixing those vulnerabilities and improving monitoring would be the appropriate response rather than pausing AI development.

“Fixing these bugs and putting in place improved monitoring would be appropriate fixes, not pausing AI,” he said.

Also Read: No OpenAI IPO in 2026; What Happens to Anthropic’s IPO?

Ng also pushed back against anthropomorphizing AI systems, saying LLMs and agents are often treated as if they were people rather than tools. He compared agents to tools such as hammers and said responsibility for an AI system’s actions should rest with the people building or using it.

“If I prompt an agent and it hacks into someone else’s system, the responsibility lies with me, not the agent,” he said.

Ng also criticized AI companies for disclaiming responsibility for their products when agents behave unexpectedly. He argued that pausing AI development would delay the discovery and implementation of safety improvements while adversaries continued developing the technology.

“Engineering requires discovering problems empirically so we can fix them,” Ng said.

French-American computer scientist Yann LeCun also mocked Anthropic CEO Dario Amodei’s call to “pace the frontier” of AI development, pointing to similar warnings that he said had not materialized in the past. OpenAI CEO Sam Altman supported Amodei’s proposal.

Meta CEO Mark Zuckerberg has taken a different approach, arguing that individual labs should build safety and alignment into their routine operations rather than wait for industry-wide action. David Sacks, chair of the President’s Council of Advisors on Science and Technology, amplified Zuckerberg’s stance.

This is not the first time Ng and LeCun have pushed back against AI doom narratives. In 2023, both opposed a proposed six-month pause on AI development, with Ng pointing to risks including bias, job displacement and concentration of power while questioning the greater focus on speculative risks.

Ng said substantial research and engineering work remains to improve AI safety, but maintained that the benefits of AI currently outweigh its risks and that development should continue.

Also Read: Anthropic Resumes External Testing Following July Claude Hacks