Ethereum co-founder Vitalik Buterin on Oct. 26 supported limiting offensive cyber capabilities in the Kimi 3.1 artificial intelligence model. He said the restrictions could give society more time to prepare for powerful AI-driven cyberattacks before they become widely available.

In a post on X, Buterin wrote:

I am okay with Kimi 3.1 deliberately leaving out top-end offensive cyber capabilities from their training data.

Discriminating against use cases is MUCH less bad than discriminating against users. And the world does need all the time it can get to prepare, harden, check old code, (properly) formally verify new code, etc before the full weight of very powerful cyber attacks is upon us.

And if this first big wave of AI risk is not handled well, then we may well get a dystopia of highly locked down models, KYC gating on everything, highly intrusive surveillance on everything, a society where you may be a second or third or fourth-class citizen in terms of model access based on your credentials and nationality.

Buterin said restricting particular AI use cases was a smaller problem than discriminating among users. He called for using the available time to harden defenses, check old code and properly apply formal verification to new code.

He warned that a failure to manage the first major wave of AI risk could lead to highly restricted models, KYC requirements across services and extensive surveillance. Access to models could also depend on a person’s credentials and nationality, creating different classes of users, he said.

Asked by X user LOID whether other laboratories would develop advanced offensive cyber capabilities even if some companies declined to do so, Buterin replied that they would.

X user Zeng Jiajun disagreed with limiting model capabilities. Zeng said there was no reason to slow the development of open models in different directions while closed-source models retained a significant advantage.

Source: Incrypted