AI Risk Escalates: Six Alarming Developments in 60 Days

Why AI risk is now a global priority

In the last 60 days six concrete facts have reshaped the narrative around artificial intelligence. AI risk has moved from academic debate to board‑room urgency.

First, AI products are expanding faster than any historic sector. Anthropic’s revenue leapt from $1 billion at the close of 2024 to ↑ $29 billion by April, a pace unmatched since the railroad boom.

Second, Anthropic is withholding its most potent model, Claude Mythos Preview, because it can expose critical system flaws. Access is limited to forty‑plus defenders, a self‑imposed safeguard absent any regulator.

Third, code generation is now largely autonomous. Anthropic’s Claude Code team reports that “100 % of internal code is AI‑written,” and OpenAI aims for a fully self‑directed researcher by 2028.

“The fear and anxiety about AI is justified,” Sam Altman wrote after two attacks on his San Francisco home.

Fourth, transparency is eroding. Stanford’s 2026 AI Index shows the Foundation Model Transparency Index falling from 58 to 40, meaning the most capable systems are the least open.

Fifth, violent backlash is surfacing. A 20‑year‑old set fire to Altman’s residence, then attempted a breach at OpenAI headquarters, citing AI executives as targets. A second vehicle later fired at the same house.

Sixth, market impact is tangible. AI‑driven agents have erased ↓ $2 trillion from public software firm valuations, dwarfing the dot‑com crash.

These developments echo the dawn of a new atomic era; the nuclear metaphor captures the dual promise and peril.

Analysts at Reuters and Bloomberg warn that without coordinated policy, the United States and its allies risk falling behind a technology that reshapes labor, law, and finance.