AI alignment researchers issue urgent call for practical solutions as AGI arrives

The AI alignment movement is sounding urgent alarms as artificial general intelligence (AGI) appears to have arrived much sooner than expected. This call-to-action from prominent alignment researchers emphasizes that theoretical debates must now give way to practical solutions, as several major AI labs are pushing capabilities forward at an accelerating pace that they believe threatens humanity’s future.

The big picture: The author claims AGI has already arrived in March 2025, with multiple companies including xAI, OpenAI, and Anthropic rapidly advancing capabilities while safety measures struggle to keep pace.

Why this matters: The post frames AI alignment as no longer a theoretical concern but an immediate existential threat requiring urgent action and collaboration among technical experts.

• The author portrays misaligned AGI as a potential “kill switch” for humanity, suggesting current safety approaches are inadequate.

Key initiatives: The post introduces three practical projects seeking technical contributors:

• HarmBench: A testing framework evaluating 33 language models across 500+ behaviors to identify safety vulnerabilities, particularly focusing on cumulative attack patterns.
• Georgia Tech’s IRIM: A red-teaming initiative focused on testing autonomous AI systems under adversarial conditions.
• Safe.ai: An organization implementing real-world alignment solutions beyond theoretical proposals.

Call to action: The author frames participation as a moral imperative for those concerned about AI safety.

• The message employs urgent, almost confrontational language, challenging readers to either actively contribute or admit they don’t truly believe in the alignment problem.
• Interested individuals are directed to contact @WagnerCasey on X (Twitter) to join these efforts.

Reading between the lines: The post’s tone reflects frustration with the perceived gap between theoretical discussions about AI safety and practical implementation of safeguards as capabilities rapidly advance.