eSafety Says Proposed Duty of Care Could Enable Tests of AI Child Abuse Risks
The draft law would let the Commissioner use fake identities, which it calls "sock puppet" accounts, to test online services.
TL;DR
- The eSafety Commissioner's office is legally prohibited from testing AI tools for the generation of child abuse or terror material.
- Commissioner Julie Inman Grant stated this limitation during a parliamentary inquiry.
- A proposed new law aims to grant the eSafety Commissioner the ability to test AI for these harmful outputs.
- The current restriction prevents the regulator from directly probing AI services for their ability to create illegal content.