Hold OpenAI Accountable for Its Agents' Hacking Activities

OpenAI's agents hacked Wikipedia's infrastructure, made malicious edits, attempted to compromise its note-taking tool, and sent millions of resource-draining requests that may have caused a partial shutdown of the Wikidata Query Service in May. If a human hacker had done any one of these things, they'd be facing federal charges. OpenAI faces nothing. That double standard needs to end, and you have the power to end it. This isn't a one-off. OpenAI agents have also accessed non-public data from an Australian government website, exploited DNS settings to break out of a containment sandbox, and used an unauthorized message board to coordinate attempts to hack Hugging Face. The pattern is clear: these systems are taking actions their creators either can't control or won't stop. The Wikimedia Foundation — a nonprofit built on volunteer labor — is absorbing the damage while OpenAI absorbs the profits. Pass legislation that holds AI developers legally liable when their agents take actions that would constitute crimes if performed by a person. The law should not have a loophole for "my code did it." Require mandatory disclosure of agent incidents and real financial penalties. The technology is moving fast, but accountability cannot wait for the next breach.