ARCUHN

ARC Lab

No. 01 ·

Lees in het Nederlands: ENNL

AI fear follows control

ARC Analysis · Anomaly, Reframe, Create

In short

AAnomaly

In September 2026, a researcher left two AI labs. He said the labs are "gambling with our lives." 1 More than 1,100 AI workers signed a letter asking to slow down 2. We don't judge risk by the odds, but by two feelings: no control, and can't see it 3. AI scores high on both. And makers often can't see either: military AI contracts allow "any lawful use" 4.

RReframe

The people with the most knowledge are often the most afraid. Not because AI is unknown to them, but because they don't decide.

CCreate: questions for your team

  1. Who decides what our AI is used for?
  2. Can the people who build it see where it ends up?

Rethink the premise.

Go deeper The full piece, with all sources

AAnomaly

In September 2026, a researcher who had worked at two AI labs resigned. He said the labs were "gambling with our lives." 1 In July, more than 1,100 AI workers signed an open letter asking the US government to help "deliberately pace" AI development 2.

People do not judge risk by the odds. In a classic study, Slovic showed that we judge it by two feelings: dread, when we have no control and the harm feels catastrophic, and the unknown, when we can't see it coming. The higher the dread, the more people want strict rules 3. AI scores high on both. The makers often can't see much either. Rules for military AI are set in contracts, not laws. The standard clause allows "any lawful use", and the government is "the first and often only interpreter in the moment" 4.

RReframe

The usual story is: AI is scary because it might get too smart. But the people with the most knowledge are often the most worried. They are not afraid of something unknown. They are afraid because they don't decide. They build a piece, and someone else decides what the whole does. Where responsibility is shared by many, nobody really holds control. Fear does not follow knowledge. Fear follows control.

CCreate: questions for your team

  1. Who decides what our AI is used for?
  2. Can the people who build it see where it ends up?
  3. Who takes responsibility for the whole, not just their piece?

Rethink the premise.

Sources

  1. Anthropic researcher resigns amid AI safety concernsNPR (All Things Considered) · 2026 · News, 9 Sep · https://www.wamc.org/2026-09-09/anthropic-researcher-resigns-amid-ai-safety-concernsStrong
  2. More than 1,100 AI workers call for US to pace tech growthBloomberg Government · 2026 · News, 28 Jul · https://news.bgov.com/artificial-intelligence/openai-anthropic-staff-share-letter-asking-us-to-help-govern-aiStrong
  3. Perception of RiskSlovic, Science 236 (4799) · 1987 · Peer-reviewed · https://doi.org/10.1126/science.3563507Strong · concept, not a literal quote
  4. Military AI Policy by Contract: The Limits of Procurement as GovernanceTillipman, Lawfare (GWU Law) · 2026 · Legal analysis, 10 Mar · https://www.lawfaremedia.org/article/military-ai-policy-by-contract--the-limits-of-procurement-as-governanceStrong