Skip to content
forward pass

The Bletchley Declaration

Twenty-eight countries including the US, China and the EU sign a statement on frontier-AI risk at the UK's summit; national safety institutes follow.

category
policy
significance
3 of 5
people
Rishi Sunak, Kamala Harris
organisations
UK Government

what had to happen · 53 events back to 1943

Every event this one built on, transitively, in order. The organism has the same path lit. Direct influences are marked.

00 · One neuron · 1

  1. 1943A logical calculus of nervous activity

I · Foundations · 11

  1. 1948A mathematical theory of communication
  2. 1949Cells that fire together wire together
  3. 1950Programming a computer for playing chess
  4. 1950Computing machinery and intelligence
  5. 1958The perceptron learns
  6. 1959Samuel's checkers program coins 'machine learning'
  7. 1960ADALINE and the least-mean-squares rule
  8. 1965Moore's law
  9. 1966ELIZA
  10. 1969Perceptrons
  11. 1970Reverse-mode automatic differentiation

W1 · The first winter · 2

  1. 1974Werbos applies backpropagation to neural networks
  2. 1980The Neocognitron

II · Connection · 3

  1. 1982The Hopfield network
  2. 1985The Boltzmann machine
  3. 1986Backpropagation

W2 · The second winter · 6

  1. 1988Temporal-difference learning
  2. 1989Q-learning
  3. 1989LeNet reads handwritten postcodes
  4. 1990Finding structure in time
  5. 1991The vanishing gradient problem
  6. 1992TD-Gammon reaches world-class backgammon

III · Statistics and data · 9

  1. 1997Long short-term memory
  2. 1998MNIST and LeNet-5
  3. 1999The first GPU
  4. 2003A neural probabilistic language model
  5. 2006Deep belief networks and the word 'deep'
  6. 2007CUDA
  7. 2009Deep learning moves to GPUs
  8. 2009ImageNet
  9. 2010Rectified linear units

IV · Deep learning · 9

  1. 2012Google Brain's network discovers cats
  2. 2012Dropout
  3. 2012AlexNet wins ImageNet
  4. 2013Deep Q-networks play Atari
  5. 2014Superintelligence
  6. 2014Attention
  7. 2014Sequence to sequence learning
  8. 2015Batch normalisation
  9. 2015Residual networks

V · Transformers · 8

  1. 2017The Asilomar AI Principles
  2. 2017Attention is all you need
  3. 2017Deep reinforcement learning from human preferences
  4. 2018GPT: generative pre-training
  5. 2019GPT-2 and the model too dangerous to release
  6. 2020Scaling laws for neural language models
  7. 2020GPT-3
  8. 2020Learning to summarise from human feedback

VI · Everyone · 4

  1. 2022InstructGPT
  2. 2022ChatGPT
  3. 2023GPT-4direct
  4. 2023'Pause Giant AI Experiments'direct

The UK's AI Safety Summit met on 1 and 2 November 2023 at Bletchley Park, where Turing had worked on the German ciphers, and its first-day statement was signed by 28 governments including the United States, China, the European Union, India and Brazil. The declaration recognised that frontier models could cause "serious, even catastrophic, harm, either deliberate or unintentional", agreed that the companies building them bore responsibility for their safety, and committed the signatories to meet again. Elon Musk interviewed Rishi Sunak on stage; Kamala Harris announced an American safety institute in a speech in London the same day.

The summit produced institutions. The UK's AI Safety Institute, later renamed the AI Security Institute, began testing frontier models before release under voluntary agreements with the laboratories; the US established its own; Japan, Singapore, Canada and others followed; and further summits met in Seoul in May 2024 and Paris in February 2025, where the American delegation declined to sign the communiqué.

Bletchley is on this timeline as the first time governments treated the technology as a matter of collective safety rather than industrial policy, and as the start of a two-year period in which that framing held before the competitive one returned.

what it led to · 5 events downstream, through 2026

Built on it directly:

  1. 2024The EU AI Act enters into forceVI

And, through them, by era:

VII · Agents · 4
  1. 2025America's AI Action Plan
  2. 2026Claude Fable 5 and the Mythos class
  3. 2026The EU delays its high-risk AI rules
  4. 2026Claude Fable 5.1

sources · 1

Trace the lineage of this event on the timeline →Back to the ledger