Obama: AI concerns “not overhyped”

Summary
In this clip from a September 18, 2026 speaking appearance broadcast by C-SPAN, former U.S. President Barack Obama shares his perspective on the acceleration of artificial intelligence. He argues that while market valuations and commercial claims may be inflated, the technology’s exponential progress and capacity for recursive learning are genuine, creating both alignment risks and immediate dangers from weaponization by bad actors.

What is shown

• [00:00] Barack Obama speaking on stage with a handheld microphone regarding AI market hype versus actual technological progress.
• [00:27] Obama explaining recursive learning and using hand gestures to illustrate the shift into a steep “hockey stick” growth curve.
• [01:12] Obama discussing sci-fi existential risk narratives versus practical goal misalignment issues.
• [02:37] Obama identifying weaponization by malicious actors as the more pressing near-term danger.

Claims & numbers

• The speaker claims the commercial benefits and company valuations around AI may be overhyped, but the underlying technology is not (the presenter says this).
• The speaker claims AI has entered a phase of recursive learning where systems can begin teaching themselves rather than waiting on human input (the presenter says this).
• The speaker claims AI development has shifted from linear progress to an exponential “hockey stick trajectory” (the presenter says this).
• The speaker claims there is a “non-zero chance” of catastrophic loss-of-control scenarios where machines set their own agendas, though he considers misalignment and malicious human misuse to be more realistic and urgent threats (the presenter says this).

Notable quotes

• [00:00] “I do not believe that this technology is overhyped. The commercial benefits of it may be overhyped... the technology itself is not overhyped.”
• [00:38] “...the machines are starting to be able to teach themselves. They don’t have to wait for humans to start figuring things out.”
• [01:44] “...there is a non-zero risk of—a non-zero chance of that happening. But that’s not actually the risk that I’m most concerned about...”

Assessment
This is a standard broadcast recording of a public speech and Q&A session, not a product demonstration or benchmark test. The presentation consists of general analytical and policy commentary on the pace and risks of frontier AI models.

Described by gemini-3.8-flash on 2026-10-18 from the video’s audio and frames.