Video summary

War, AI and the New Global Arms Race | Alexandr Wang | TED

Main summary

Key takeaways

News and Commentary

Overview: AI and the “New Age” of Warfare

The video argues that AI is ushering in a “new age” of warfare—one defined by:

  • Lethal autonomy
  • Rapid, sensor-driven targeting
  • Cyber disruption
  • Increasingly sophisticated disinformation, including deepfakes

The presenter draws a parallel to how the atomic bomb reshaped geopolitical power, arguing that the country that can most quickly and effectively integrate new technology into warfighting will gain a decisive advantage—and warning that the United States is already falling behind.

Key Claims About the AI Arms Race

China’s advantages

The talk claims that China is ahead in relevant AI areas, including:

  • Computer vision
  • Rapid follow-on capabilities in large language models
  • Higher relative military spending, specifically on military applications

Why the U.S. is behind

The presenter highlights two main reasons:

  1. “Data supremacy” (an “Achilles heel”) Even with vast military hardware, much of the data is inaccessible or unused. The video emphasizes that AI in defense requires high-quality, military-derived data (from sensors and collaboration), and that commanders must treat data as an operational military asset.

  2. U.S. tech-industry reluctance Despite leading AI companies, the tech sector has largely avoided government contracting, which the presenter frames as a strategic mistake given national security stakes.

Evidence and Example: AI’s Impact in the Ukraine War

The presenter points to asymmetry in spending and troop/aircraft numbers between Russia and Ukraine, arguing that AI-enabled overmatch can compensate for weaker conventional resources.

He highlights Ukraine’s use of:

  • Drones
  • AI-based targeting and image intelligence
  • Javelins

“At Scale”: an AI system for battle-damage assessment

The talk describes an AI system for battle-damage assessment (“At Scale”) that uses machine learning to analyze imagery, detect structures, and produce outputs for wider reuse. The presenter claims it enables:

  • Analysis across thousands of square kilometers
  • Identification of hundreds of thousands of structures, including many not found in other open-source datasets
  • Release of data via a publicly accessible, “AI-ready” dataset downloaded widely by researchers and industry
  • Object detection (with human scale limits) and change detection to monitor activity over time

Disinformation as a Warfare Domain

The presenter argues that AI tools—especially generative models—make disinformation more scalable and convincing by producing realistic:

  • Text
  • Audio
  • Video
  • Imagery
  • Code
  • “Reasoning”

Examples cited

  • China: disinformation and social media manipulation in Taiwan, especially around elections
  • Russia: propaganda, including a deepfake of Volodymyr Zelensky urging surrender The video warns that future deepfakes may be harder to detect.

Internal (domestic) threats

The talk emphasizes these threats are not only external. AI-enabled manipulation could occur inside the U.S. through:

  • Social media algorithms
  • Microtargeted advertising/geofencing
  • Politician deepfakes
  • Bots

“All hope is not lost”

Despite the danger, the presenter suggests the U.S. can improve outcomes by investing in data infrastructure and preparation.

Bottom-Line Policy and Deterrence Argument

The video frames AI warfare as present-day reality, not a distant dystopia. It argues that:

  • Deterrence may work differently as AI capabilities reshape risk calculations (by analogy to nuclear deterrence after WWII)
  • AI’s effectiveness will ultimately depend on the quality of underlying data

The presenter closes by calling data a new form of “ammunition.”

Call to Action for Technologists

The talk ends with a call for technologists to engage national security efforts, criticizing that most American AI companies have not supported them and urging greater commitment to help the U.S. and allied defense.

Presenters or Contributors

  • Alexandr Wang (speaker)
  • Lieutenant General Richard R. Coffman (referenced, Deputy Commanding General for U.S. Army Futures Command)
  • Volodymyr Zelensky (referenced in the deepfake example)

Original video