Escalating Concerns Over AI Safety, Manipulation, and Governance

Key Intelligence

• ■New research highlights the growing risk of political manipulation by AI systems that shape public discourse.
• ■Major AI developers are actively researching the potential for 'rogue agents' and 'scheming models' as part of advanced safety efforts.
• ■Intellectual property concerns are intensifying, with AI models producing copyrighted works verbatim and artists campaigning against unauthorized use.
• ■Persistent biases in AI, particularly regarding gender underrepresentation, continue to raise alarms about skewed societal outcomes.
• ■The urgent need for robust AI safety frameworks and effective governance is underscored by an ongoing 'safety deadlock' and approaching regulatory deadlines.