Interview
“Russia and Ukraine are inventing the future of war”: former Google CEO Eric Schmidt | The Economist
- Future warfare is predicted to be significantly more destructive and dystopian than current media portrayals suggest.
- Russia's "Rusvet" satellite network is expected to be operational within approximately two years, potentially becoming unblockable and enabling targeting capabilities inside Ukraine.
- Russia plans to deploy up to 80,000 jet-powered Shahed drones against Kyiv and other targets between now and the end of December.
- Operationalization of the Rusvet network is anticipated to significantly increase Ukraine's ability to inflict damage on Russian oil infrastructure.
- Weapon systems are expected to evolve to include a final autonomous segment that hits targets precisely without human intervention in the final moment.
- Large language models are projected to eventually reason around objectives to determine optimal asset positioning for achieving strategic goals.
- Future end systems will anticipate losses and provide generals with recommended solutions to military problems.
- Unchecked AI autonomy poses a risk of executing massive war crimes or attacking uninvolved nations due to a lack of human judgment.
- The trajectory toward full AI autonomy in warfare carries a risk of a "Dr. Strangelove situation" if established rules are not implemented.
- The rapid rate of innovation in the current conflict creates an uncontrollable logic of escalation that may outpace human debate.
- U.S. military resistance to adopting cheaper, more effective technologies is driven by loss aversion regarding personnel casualties.
- Over time, U.S. military leadership is expected to converge on the necessity of new drone-based architectures to protect national interests.
- A future conflict in the Taiwan Strait is predicted to resemble a drone war rather than a ship war due to the vulnerability of large vessels.
- A pause on AI development in the U.S. is unlikely due to competitive incentives that favor continuous advancement over halting progress.
- A "mutual AI destruction" protocol, allowing nations to remotely power down opposing data centers, is suggested as a potential but far-fetched solution to the pause issue.
- Future AI systems will remain under human control and alignment despite their operational speed.