OpenAI Expands Bug Bounty to Cover AI Abuse and 'Safety' Concerns (infosecurity-magazine.com)

• Inclusion of model safety and misuse under bounty scope.
• Targeting adversarial manipulation and systemic ethical failures.
• Crowdsourcing identification of non-traditional AI vulnerabilities.

"OpenAI has officially expanded its bug bounty program to incorporate non-traditional security vulnerabilities, specifically targeting AI model safety and misuse. By incentivizing researchers to identify systemic risks and harmful outputs, the initiative aims to build a more resilient framework for mitigating adversarial manipulation and ethical failures in large-scale language models."