* Originally in: SFSciFiRea

'That branch of AI is lethal. We've got to do something about that' Neil deGrasse Tyson wants to ban AI superintelligence

Date:
Tue, 24 Mar 2026 13:13:33 +0000

Description:
Neil deGrasse Tysons call to ban AI superintelligence underscores rising
fears that future AI could outpace human control and force global action

FULL STORY
Neil deGrasse Tyson is not usually the
person in the room calling for a global ban on anything. He is better known
for explaining black holes with a smile than for advocating international treaties.

But in a recent talk that has been circulating widely online see below the astrophysicist delivered a stark warning about artificial intelligence that sounded less like a science lecture and more like a line from a disaster
movie. "That branch of AI is lethal," he said. "We've got to do something
about that. Nobody should build it."

The branch he is referring to is artificial superintelligence, a hypothetical future form of AI that would surpass human intelligence across nearly all domains. For Tyson, the concern is not incremental improvements in chatbots
or image generators. It is the possibility of something far more powerful, something that could outthink, outmaneuver, and potentially outlast its creators.

Most peoples daily experience of AI is a chatbot drafting emails, a phone organizing photos, or a navigation app rerouting around traffic. Tysons warning, though, taps into a growing debate that has moved from academic
papers into mainstream conversation.

The idea of banning superintelligence is not new. Researchers and public figures have been discussing it for years, often framing it as a precaution against an intelligence explosion, where AI systems rapidly improve
themselves beyond human control.

Some proponents argue that once such systems exist, it may be impossible to contain them or align them with human values. The counterargument is that
these fears are speculative and risk slowing down beneficial innovation.

Tysons contribution stands out for its clarity and a suggestion of global cooperation on a ban.

Everyone needs to agree to that by treaty," he said. "Treaties are not
perfect, but theyre the best we have as humans.

International treaties are one of the few mechanisms humanity has for
managing existential risks. Nuclear weapons, chemical weapons, and even ozone-depleting substances have all been subject to global agreements. The logic is simple, even if the execution is not.

If a technology is too dangerous for any one country to handle alone, then it becomes everyones problem. But AI is software, not a bomb, and software has a way of slipping across borders. AI proliferation and fear High-profile voices have continuously warned that AI could be dangerous enough to warrant global intervention, even as the technology becomes ubiquitous. You might use AI to plan a weekend trip or summarize a meeting, all while hearing that the same underlying technology could one day become uncontrollable.

Tysons call for a treaty does not resolve that tension. If anything, it sharpens it. As regulation has often lagged behind innovation, his call for a treaty when superintelligence seems purely theoretical isn't absurd. Usually, by the time governments act, a technology has already become widespread.

AI may be different in that its potential risks are being discussed before
its most advanced forms exist. That creates an opportunity, but also a
dilemma. Acting too early could stifle progress. Acting too late could make control impossible.

What Tyson is suggesting is that the answer should not be left to chance. But like most collective decisions, it is likely to be messy, contested, and far from unanimous.

Link to news story: https://www.techradar.com/ai-platforms-assistants/that-branch-of-ai-is-lethal- weve-got-to-do-something-about-that-neil-degrasse-tyson-wants-to-ban-ai-superi ntelligence

$$
--- SBBSecho 3.28-Linux
* Origin: Capitol City Online (1:2320/107)