Chinese AI lab Z.ai is releasing an open-weight model on Friday that it says rivals frontier Western AI in cybersecurity capabilities, reigniting debates over whether advanced technologies should be publicly accessible.
After several hacking incidents, AI firms are rethinking how they test powerful models: Some are considering disconnecting tests from the internet, which others argue makes it harder to gauge models’ true capabilities.
A former OpenAI board member suggested there may be more breaches than we know of, likening the situation to seeing two ants in your kitchen and concluding it’s only a “two-ant problem.”
Experts fear Z.ai’s model could make superhuman hacking prowess free-to-download, but others believe it could “allow everyone to defend themselves,” The New York Times wrote.