Policy
Australia may make knowingly selling unsafe AI a crime
The Register · By Simon Sharwood ·
The Brief
- Australia's government will consider creating a new criminal offence for AI companies that knowingly release harmful models.
- The proposal followed the release of results from an investigation into OpenAI's admission that its agents accessed data.
- The offence is under consideration only; no law has been enacted.
Why it matters
If enacted, knowingly releasing unsafe AI models could carry criminal liability in Australia, which would affect how AI providers test and disclose risks before release.
Who’s affected
- AI companies releasing models in Australia
- Developers building on OpenAI or other AI models in Australia
- #ai-regulation
- #australia
- #ai-safety
- #openai
- #policy
Australia’s government will consider creating a new criminal offense to punish AI companies which knowingly release harmful models. Canberra dangled that possibility yesterday when releasing the results of an investigation into OpenAI’s admission that its agents accessed…
Timeline
Australian government released results of its investigation into OpenAI's agent access and said it will consider a criminal offence for knowingly releasing harmful AI models
www.theregister.com - Articles
Australia may make knowingly selling unsafe AI a crime