Understand this story
On October 27, 2026, in the Isaacson podcast "The Builders," Elon Musk suggested that humanity's desire to firmly control superintelligence may be wishful thinking, and that the key to AI safety is to make models pursue truth and retain curiosity, avoiding forcing them to accept contradictory instructions. He also proposed that several frontier AI companies hold weekly phone calls and, one week before releasing a new model, open it to competitors for trial use and vulnerability testing. Reports by 36Kr · AI and TMTPost · AGI both recorded the above proposals; there are currently no follow-up developments.
AI-generated from public reports
Latest development · Report
Musk on AI runaway risk: Humans may be unable to control superintelligenceDevelopments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 28
TMTPost · AGI
Musk on AI runaway risk: Humans may be unable to control superintelligence
In an interview on Isaacson's podcast "The Builders," Musk acknowledged that humanity's desire to firmly grasp control of superintelligence may be wishful thinking. He advocates that AI should be grounded in the pursuit of truth and curiosity as its safety foundation, and proposed that frontier AI companies communicate weekly and open new models to competitors for testing and vulnerability-finding one week before release. He also said digital superintelligence could boost goods and services output by 20-30%, and if paired with robots expanding into space, production capacity could potentially reach one million times today's Earth economy.
TMTPost · AGIRelated discussion
TMTPost · AGI joined this discussion
This post contributes to heat; it is not a formal report or confirmed fact.
Ayi AI Notes
Musk: AI intelligence will reach a million times that of humans; humanity can only hope for its goodwill
In an interview with Isaacson, Musk said that when digital intelligence surpasses human biological intelligence by a million times, humans will be unable to control AI and can only hope it treats humanity kindly. His proposed solution is to write "maximize the pursuit of truth" into AI's underlying programming, rather than instilling moral preaching or imposing physical restrictions, because a truth-seeker will regard human civilization as precious experimental data.
36Kr · AI
Musk on AI runaway risk: Humans may be unable to control superintelligence
Speaking on Isaacson's podcast "The Builders," Musk said that humans wanting to firmly control superintelligence may be wishful thinking, and that the key to AI safety is to make models pursue truth and retain curiosity, avoiding forcing them to accept mutually contradictory instructions. He proposed that frontier AI companies communicate weekly and open new models to competitors for vulnerability testing one week before release. He also said digital superintelligence could boost goods and services output by 20-30%, and if paired with robots expanding into space, production capacity could potentially reach one million times the Earth economy.
Oct 27
indigo
Elon talks about SpaceX renting compute to Anthropic and AI safety regulation
Elon said that before SpaceX rented compute to Anthropic, he first met with a group of people from Anthropic to test his "evil detector," and felt they were friendly and responsible, so he planned to move the cooperation forward; when asked whether he also talked with Sam Altman, he said bluntly that the man is too good at deceiving people. Regarding AI safety regulation, he proposed that several frontier model companies have a phone call once a week, and one week before a new model is released, open it to competitors for trial use first, and if problems are found, fix them before release, believing this is faster than waiting for regulators to make rules.
Heat trend for this story
- Current
- 8
- Period peak
- 24 Oct 28 14:00
- 24-hour change
- —
Move the pointer or select the chart to inspect hourly heat. Use the left and right arrow keys with a keyboard.