Sam Altman on AI’s future: human control, accountability and international standards
What Sam Altman’s September 23, 2026 UN Security Council remarks on human oversight, accountability and international AI standards mean for businesses.
On September 23, 2026, OpenAI CEO Sam Altman addressed the UN Security Council. He described both the opportunities of AI and the danger of increasingly autonomous systems making decisions people cannot understand or control. For businesses, the speech offers a useful lens on AI as part of operations that people must still oversee.
AI as everyday business infrastructure
Companies use AI employees to answer recurring questions, find information in company materials and prepare customer requests for a specialist. Once these systems sit inside daily workflows, answer quality depends on current knowledge, permissions and review by staff.
Altman did not argue that every task should be automated. He stressed that some things cannot and should not be automated. A business should decide which tasks the system supports and which decisions stay with its people.
Design human oversight into the workflow
Keeping people at the center of AI decision making was a core principle of the remarks. In practice, a company can define the AI employee’s role, the information it may use, the questions it may answer and the point at which it hands a request to a person.
Oversight also requires a way to correct a bad answer and intervene before an error affects a customer. That calls for current source material, tests of ordinary and difficult cases, and a clear owner for escalated questions.
Accountability and transparency
Altman called for complementary national and international standards to measure capabilities, assess risks and safeguards, and preserve meaningful human oversight. He also discussed incident reporting. These are proposals for cooperation, not standards already enacted for every business.
At company level, customers and staff should know when AI is answering, what information supports the answer and who owns the final decision. A system should not present a guess about price, availability or terms as a verified fact.
A practical example in NEXSTA
NEXSTA illustrates one way to put AI to work in a business. A company sets the AI employee’s role, supplies its product and service knowledge, adjusts settings and operating rules, tests answers before publication and can pass a question to a person.
The company remains responsible for the accuracy of its materials and customer decisions. Automation can help staff with routine exchanges while leaving consequential judgment under human control.
For businesses, response speed is only one measure of an AI deployment. Control, verifiable knowledge, clear limits and a person able to intervene matter just as much.