Sam Altman Says OpenAI Opposes Ceding Control of the Future to Superintelligence and That This Should Ultimately Constrain AI Progress
People should always be in control of the future. OpenAI is against the notion that it would be acceptable for people to be relegated to a role of doing things that “AI systems think makes sense to reward in humans”. People should not cede control of the future to superintelligence, and that requirement should ultimately constrain AI progress.
Context
The head of the leading AI lab publicly saying human control should limit how far AI progress goes is a strong statement of policy that rivals, regulators and his own staff can hold him to.
Debates about superintelligence often turn on whether very capable AI systems should be allowed to make decisions for people at large scale. Sam Altman (@sama) is the CEO of OpenAI. Keras creator François Chollet (@fchollet) quote posted the statement to agree with it.
Deep dive
Sam Altman (@sama), the CEO of OpenAI, posted on the evening of Nov 3 ET that people should always be in control of the future.
In the post he says OpenAI is against the notion that it would be acceptable for people to be relegated to a role of doing things that AI systems think makes sense to reward in humans. The phrase about rewarding humans appears in quotation marks, which suggests he is responding to an argument made elsewhere, but the post does not say who made it or link to it.
He then goes further than a general statement of values. He says people should not cede control of the future to superintelligence, and that this requirement should ultimately constrain AI progress. In other words, he presents human control as a limit on how fast or how far AI development should go, not just as a design goal.
The post passed 2.7 million views within a few hours and drew more than 900 quote posts. François Chollet (@fchollet), the creator of Keras, quote posted it with a co sign, saying the only future worth building is one where AI remains a tool in human hands used to elevate the human condition. Many other replies questioned how the statement fits with OpenAI's push to build ever more capable systems.
The post is a short statement of position, not a policy document. It does not describe any concrete mechanism, such as specific limits, oversight bodies or tests, that OpenAI would use to keep people in control.
What to expect next
Watch for whether OpenAI turns the statement into something concrete, such as changes to its safety framework, published limits on autonomous systems or commitments in its policy work. Other lab leaders may be asked to say whether they agree. Critics will likely test the claim against OpenAI's product plans for more autonomous agents. A key open question is who decides when AI progress should be constrained, and how.