The Case for Mandatory AI Monitoring & Incident Reporting

Transparency into AI companies and the impacts their products are having is key to informing effective regulatory responses. This presentation will survey how other jurisdictions (including the EU, California, New York, and Vietnam) have instituted AI monitoring and reporting obligations to provide visibility on AI-related incidents and near misses (including AI-enabled harms (such as from cyber or biological misuse) and incidents arising from issues within the AI model itself (such as misalignment or loss of control). It will then explain why Australia should institute a similar mechanism, so that our policymakers and regulators can be fully informed of the rapidly evolving risk environment.