▼AI Offensive Cyber Capability and State Doctrine

Score history

Daily conviction score, 0 to 100. Higher means the thesis is more strongly corroborated.

Now 30 · -2 since Jul 6 · ranged 30 to 32

Showing the last few days. Unlock full score history.

Why the conviction moved

• Jun 27
  Strengthened +4

  Anthropic maintains a Frontier Red Team probing models for dangerous capabilities as states increasingly treat frontier AI as an offensive cyber instrument.

• Jun 27
  Strengthened +3

  The U.S. eased its block specifically to let select cyber defenders keep using Anthropic's Mythos 5 model, underscoring that frontier AI is being managed as a national cyber-security instrument.

• Jun 24
  Strengthened +6

  The Five Eyes intelligence alliance warned of AI offensive cyberattack capability the same day OpenAI shipped a security-tuned model with a lowered refusal rate for security work, showing states and labs converging on AI as a cyber instrument.

Showing the last 2 days. Unlock the full record.

Source trail

• Supporting · June 27, 2026

  Anthropic's Frontier Red Team and the offensive-cyber question

  Anthropic maintains a Frontier Red Team probing models for dangerous capabilities as states increasingly treat frontier AI as an offensive cyber instrument.

  Anthropic Research

• Supporting · June 27, 2026

  U.S. Eases Anthropic Block, Allowing Mythos 5 for Select Cyber Defenders

  The U.S. eased its block specifically to let select cyber defenders keep using Anthropic's Mythos 5 model, underscoring that frontier AI is being managed as a national cyber-security instrument.

  AI Post (Telegram)