Interview
“Russia and Ukraine are inventing the future of war”: former Google CEO Eric Schmidt | The Economist
- Current态势 (Current态势): Russia is shifting toward a "total war" strategy, rapidly closing the innovation gap with Ukraine despite Ukraine's initial advantage.
- Resource Disparity: While Ukraine leads in innovation speed, Russia possesses vastly superior resources and is accelerating its improvement trajectory at a similar slope.
- Critical Vulnerabilities (Ukraine): Ukraine currently lacks effective defenses against five specific Russian capabilities:
- Ballistic missiles (Patriot systems are in critically short supply).
- Jet-propelled Shahed drones (which fly faster, higher, and harder to intercept).
- Large-scale mobilization of human infantry.
- A new satellite network (Rusvet) expected to launch next year.
- Powerful independently targeted glide bombs destroying front-line positions.
- Specific Threat Data: Reports indicate plans to deploy up to 80,000 jet-propelled Shaheds against Kyiv and other Ukrainian towns by the end of December.
- Russian Counter-Satellite (Rusvet):
- Russia is building a mesh network using Chinese parts to replace Starlink capabilities in Ukraine.
- This network enables 8K video relay and real-time targeting; early evidence suggests it is functional, though fully operational capability is estimated to be one year away.
- Once operational, this network would be "more or less unblockable," allowing Russia to hunt targets anywhere in Ukraine.
- Starlink Controversy:
- Ukraine seeks permission to use Starlink for offensive strikes inside Russian territory (e.g., targeting oil infrastructure).
- Schmidt clarifies that this is a political decision, not a technical limitation; the U.S. President could order SpaceX to enable service over Russia.
- SpaceX currently restricts usage over Russia as a policy, not a technical constraint.
- AI and War Definitions:
- Current "autonomous" features are primarily machine vision for terminal guidance, not true AI; they are often disabled when human judgment proves more reliable against jamming.
- True AI involves "recursive self-improvement" and is just beginning to be used in war planning (e.g., automating troop movement).
- Schmidt envisions a future where Large Language Models (LLMs) analyze all data to recommend strategic solutions to generals.
- AI Safety and Ethics:
- Schmidt argues that full autonomous decision-making (e.g., "attack this country and tell me when you're done") is unacceptable due to the high risk of war crimes and lack of judgment.
- He asserts that "human-in-the-loop" is a mandatory boundary for all military AI decisions, regardless of an opponent's adherence to international law.
- Historical precedent (Nuremberg trials) suggests legal accountability will still apply even in total war scenarios.
- U.S. and European Military Adaptation:
- The U.S. military is described as "incredibly change-resistant" due to political lobbying, bureaucracy, and loss aversion (fear of casualties).
- Schmidt suggests the U.S. must allocate new funds to innovative, cheaper systems (drones, AI) while protecting traditional budgets to allow adoption.
- European militaries suffer from similar issues, often operating as "jobs programs" rather than combat-optimized forces, and are currently facing a missile defense shortage.
- Ukraine as a Tech Model:
- European nations are encouraged to license drone technology from Ukraine rather than building it domestically to avoid delays.
- Ukrainian drone brigades are led by civilian tech leaders who have replaced Soviet-era generals.
- Work ethic in Ukrainian drone factories is described as extreme, with workers (often exempt from draft) laboring 12+ hours/day, six days a week.
- China Strategic Assessment:
- Schmidt believes the U.S. is in a better position than perceived because China has not fought a war in 50 years and lacks experience in subtle, real-time tactical warfare.
- He anticipates future Pacific conflicts (e.g., Taiwan Strait) will be dominated by drone ships rather than traditional aircraft carriers.
- China's dominance in consumer drones (DJI) poses a risk, but its lack of military combat experience remains a critical handicap.
- AI Regulation and US-China Relations:
- Schmidt opposes a "pause" on AI development due to incompatible incentives and the impossibility of verification by international bodies.
- He proposes a "no surprises" treaty as the first priority to prevent accidental escalations (e.g., AI errors triggering nuclear responses).
- A concept of "mutual AI destruction" was suggested as a theoretical check-and-balance, where systems can be remotely shut down by the opposing party.
- Optimism on Control:
- Despite the alarming pace of innovation, Schmidt maintains optimism that AI can remain under human alignment and control, provided diplomacy manages state-level risks.
- He notes that while escalation is dangerous, the fundamental ability to "pull the plug" remains with human operators.