Training will rely on specialist input rather than general LLM data, and the system is currently in testing with no release date announced.

A New Zealand startup, ThroughLine, is collaborating with OpenAI, Anthropic, and Google to build an intervention tool that steers users showing violent extremist tendencies toward crisis support, including access to human counselors and a global network of helplines.

The project is guided by The Christchurch Call initiative, embedding ethical and counterterrorism considerations and ensuring alignment with global anti-extremism goals.

Key implementation questions remain, such as when authorities should be alerted in high-risk cases and how to prevent escalation within the system.

Public reaction is split between optimism for AI-enabled safety and concerns about privacy, surveillance, and misidentification.

Founder Elliot Taylor emphasizes expanding safety coverage without indiscriminately involving governments, given the sensitivity of distress signals.

Taylor warns that bypassing safeguards could push users to unregulated platforms, potentially increasing risk.

Industry responses are cautiously optimistic about proactive moderation and support networks, but effectiveness depends on robust follow-up and proper escalation protocols.

Applications could extend to content moderators, parents, caregivers, and gaming moderators, with wider implications for platform intervention strategies.

The initiative aims to provide real-time support through consultations with counterterrorism experts, without relying on training data from base LLMs.

There is no set release date; plans include exploring follow-up mechanisms and potential alerts to authorities while weighing risks of escalation and the need for non-punitive intervention.

Broader AI safety implications emphasize responsible deployment, human oversight, and cross-border collaboration that could influence global AI governance in handling extremism and online safety.