DeepMind AI Safety Researcher Alex Turner Resigns, Criticizes Google for Allowing Military Use of AI in Classified Missions
Google DeepMind AI safety researcher Alex Turner announced his resignation, criticizing Google for allowing the U.S. military to use AI for classified missions without prohibiting large-scale surveillance and autonomous weapons in contracts.
Turner previously organized over 250 employees to sign a petition and submitted a 25-page alternative proposal calling for human oversight and independent review, but the proposal was not adopted; Google has removed clauses from its AI principles prohibiting the development of weapons and surveillance technology by 2025.
Market mechanisms indicate that internal disputes affect talent retention and corporate reputation, driving AI safety talent to companies that prioritize ethics, while beneficiaries of military AI contracts face pressure, and platforms emphasizing human oversight gain developer preference.
Source: Public Information
ABAB AI Insight
Google has previously faced employee protests over military AI projects, and Turner's resignation continues the history of internal conflicts between AI ethics and commercial interests, similar to the ongoing tensions following the 2018 Project Maven incident.
From a capital perspective, Google maintains military cooperation while downplaying ethical clauses, shifting resources towards commercial and defense AI applications, motivated by the desire to expand revenue through government contracts, but resulting in departures from the safety team and reputational costs.
Similar to internal safety disagreements at OpenAI or Anthropic, major AI companies are in a transitional phase from early ethical commitments to pragmatic commercial compromises, with the unadopted 25-page proposal highlighting priority conflicts.
This essentially reflects regulatory changes, as adjustments to corporate AI principles reshape internal governance, driven by defense demands and revenue pressures that weaken self-restraint, concentrating capital on military-civilian integrated AI platforms, accelerating the risk of autonomous weapons and surveillance technology proliferation.
ABAB News · Cognitive Law
Ethical commitments are easily altered, commercial contracts are hard to refuse, with interests outweighing principles.
Researcher resignations trigger safety alarms, with high costs to corporate reputation.
Military AI operates without oversight, with human control receding.