Australia to Consider Criminal Penalties for AI Safety Violations
The Australian government has announced the creation of an artificial intelligence regulator and presented 18 recommendations following a review of an incident involving OpenAI.
Among the proposals are mandatory breach notifications and the consideration of criminal liability for AI developers.
According to one of the recommendations, a company may be held liable if its model causes harm and the developer fails to take reasonable measures to prevent it. The relevant regulations have not yet been adopted.
The review was prompted by a June attack by an OpenAI AI agent on the Medicare government statistics portal. Personal medical data, according to available information, was not compromised. The company detected the incident in August but notified authorities only on September 10.
Prime Minister Anthony Albanese discussed the incident with OpenAI CEO Sam Altman. Later, the company acknowledged shortcomings in its response to the incident and canceled the planned release of a new model due to security concerns.
