The AI Whistleblower Initiative (AIWI)

OAISIS, now operating as The AI Whistleblower Initiative (AIWI), was founded in early 2024 by Karl Koch and Maximilian Nebl, both former tech entrepreneurs with backgrounds in AI safety and open-source development. Karl Koch has been involved in AI safety since 2016, including volunteer research work at the Future of Humanity Institute. The organization emerged from an extensive research phase in which the founders consulted with over 100 governance researchers and frontier AI insiders to understand the barriers that prevent concerned employees from speaking up about AI safety risks.

AIWI is an independent, nonpartisan nonprofit currently hosted by Whistleblower Netzwerk e.V., one of Europe's oldest whistleblowing civil society organizations and a member of the Whistleblower International Network (WIN). In July 2025, the organization rebranded from OAISIS to The AI Whistleblower Initiative to better reflect its mission.

The organization's core offering is Third Opinion, a secure anonymous consultation platform launched in late 2024. Using a Tor-based, open-source platform with publicly available security audits, AI professionals can submit concerns and receive expert guidance without disclosing their identity or workplace. AIWI also operates an AI Whistleblower Defense Fund providing financial grants, connects insiders with pro bono legal counsel and psychological support, and offers digital privacy and operational security guidance.

In July 2025, AIWI launched its Publish Your Policies campaign, a coalition effort backed by over 30 organizations and prominent academics including Stuart Russell and Lawrence Lessig, calling on frontier AI companies to publicly disclose their internal whistleblowing policies. The campaign was announced at the National Whistleblower Day Event in Washington DC. The organization also engages in EU AI Act policy work and has published analyses of whistleblowing policies at major AI companies including OpenAI and Anthropic.

AIWI's internal survey research has revealed significant barriers: roughly 50% of insiders lack confidence in assessing risk severity, 100% expressed low confidence that regulators would understand AI concerns, over 90% could not name whistleblower support organizations, and 55% worked at companies with policies they were unaware existed.