ARCUHN

ARC Lab

No. 02 ·

Lees in het Nederlands: ENNL

AI turned 70

ARC Analysis · Anomaly, Reframe, Create

In short

AAnomaly

This autumn, Dartmouth celebrated 70 years of AI. In 1956 the question there was: can a machine do what a human can? In 2026 it is: what must remain human? 1 The warnings are as old as AI. In 1960, Wiener wrote that we must be sure the purpose we put into the machine is the purpose we really want 4. In 1966, people already trusted a simple chatbot with their secrets 5.

RReframe

AI is not new. What is new is the scale. And the more we trust AI, the less we think critically 7.

CCreate: questions for your team

  1. What do you hand over to AI that you should still check yourself?

Rethink the premise.

Go deeper The full piece, with all sources

AAnomaly

At the end of October 2026, Dartmouth College celebrated 70 years of AI with a conference called "What Must Remain Human?". The field was born there in 1956 1. Then the question was: can a machine do what a human can? Seventy years later, the question has turned around.

The promise is old. The 1956 team thought one summer could bring "a significant advance" 2. Every decade said "almost", and twice the money stopped when the promises did not come true 3. The warnings are just as old. In 1960, Norbert Wiener wrote: "we had better be quite sure that the purpose put into the machine is the purpose which we really desire." 4 In 1966, Joseph Weizenbaum built ELIZA, a chatbot that only turned your words into questions. His secretary knew how simple it was, and still asked him to leave the room so she could talk to it in private 5.

RReframe

The warnings were there. The trust was there. Maybe the promise made us stop listening. And by the time we saw what AI could do, changing course had become hard: when change is easy, you can't see the need, and when you see the need, change is expensive 6. AI is not new. What is new is the scale. A 2025 study of 319 knowledge workers found that the more people trust AI, the less they think critically 7. And what we automate, we stop practising 8.

CCreate: questions for your team

  1. Think first, then ask AI.
  2. Check what you hand over.
  3. Keep practising what AI does for you.
  4. Put a name on every decision.
  5. Which of these does your team already do?

Rethink the premise.

Sources

  1. AI@70: What Must Remain Human?Dartmouth College · 2026 · Primary · https://ai.dartmouth.eduStrong
  2. A Proposal for the Dartmouth Summer Research Project on Artificial IntelligenceMcCarthy, Minsky, Rochester & Shannon · 1955 · Primary · https://ojs.aaai.org/index.php/aimagazine/article/view/1904/1802Strong
  3. Artificial Intelligence: A General Survey (Lighthill report)Science Research Council (UK) · 1973 · Government report · https://www.aiai.ed.ac.uk/events/lighthill1973/lighthill.pdfStrong
  4. Some Moral and Technical Consequences of AutomationWiener, Science 131 (3410) · 1960 · Peer-reviewed · https://doi.org/10.1126/science.131.3410.1355Strong
  5. ELIZA / Computer Power and Human ReasonWeizenbaum, Communications of the ACM 9(1) · 1966 · Primary + book (1976) · https://doi.org/10.1145/365153.365168Strong · story told by Weizenbaum himself
  6. The Social Control of TechnologyCollingridge · 1980 · BookMedium · paraphrased, literal quote not checked
  7. The Impact of Generative AI on Critical ThinkingLee e.a., CHI '25 (Microsoft Research & CMU) · 2025 · 319 knowledge workers, 936 examples · https://doi.org/10.1145/3706598.3713778Strong · peer-reviewed
  8. Ironies of AutomationBainbridge, Automatica 19 (6) · 1983 · Peer-reviewed · https://doi.org/10.1016/0005-1098(83)90046-8Strong