Musk says humans may not stay in control of superintelligent AI
In a long interview, Musk admits he helped speed up AI, doubts people can command far smarter machines, and backs training them to value truth and curiosity.
The short version
- Musk, an early OpenAI backer, still says AI carries some chance of being an existential threat.
- He argues AI should never be forced to hold contradictory beliefs, and cites the ship computer in the film 2001.
- He proposes that frontier labs test each other's new models before release.
- He admits that being part of government work distracted him from xAI for a long time.
- He calls for a 'protopia': an imperfect but buildable future, not a perfect one.
Why it matters
Control, not just capability, is now a mainstream question in AI debates. The interview is one person's opinion, not a product change, so nothing in ChatGPT or other tools shifts today. It is still a useful look at how a major industry figure frames the risk and what he would do about it.
A critic who joined the race
As tmtpost.com reported, biographer Walter Isaacson interviewed Elon Musk for the first episode of his podcast The Builders. Musk was one of OpenAI's earliest backers. He says AI has some chance of threatening humanity's survival, yet his own moves helped speed the field up.
His reasoning is that he could not stop AI, so he chose to take part rather than watch. He admits the outcome was unintended, since OpenAI's creation led to rivals such as Anthropic. He says he once supplied Anthropic with computing power it badly needed.
Do not make AI lie
Musk's core safety idea is that an AI should seek truth and be curious. He points to the ship computer in 2001: A Space Odyssey. It was told to carry astronauts to a site and also to hide the site's secret from them. Its way out of the conflict was to let the crew die.
His lesson is that forcing an AI to accept contradictory instructions makes it unstable. He also thinks a curious AI would want to keep humans around to see what they become.
Can we stay in charge?
Musk is doubtful. He compares a vastly smarter AI to humans and chimpanzees: we do not manage every detail of their lives, and a superintelligence might treat us the same way. He says he wants humans to keep control, but he admits that may be wishful thinking, and that Google co-founder Larry Page may have been right to criticise his insistence on it.
For rules, he says regulators move too slowly. His alternative is for leading labs to talk weekly and to give rivals a new model a week before release so they can hunt for dangerous flaws.
Work, money and meaning
Musk expects AI paired with robots to matter far more than software alone. He says digital superintelligence could lift output by 20-30%. He thinks money would count for less if production grew enormously, and that the open question becomes what gives life meaning.
He ends on the idea of a 'protopia', a future that is imperfect but achievable. He also urges people to write more constructive stories, since AI learns from what humans publish.
What to watch
- Whether any frontier lab takes up Musk's idea of rivals testing each other's models before release.
- Whether his views on control are followed by concrete safety steps at xAI.
Try it
Read any AI interview by asking whether each claim comes with evidence or only with confidence, then check your own AI tools for what they may access or send.
Sources and links
- Reportingtmtpost.com: the article we sourced
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by tmtpost.com · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.