PauseAI protest @ Google DeepMind - May 19th - 22nd

Please read the PauseAI Protestors’ Code of Conduct here before attending this protest.


  • Lawful protest outside Google DeepMind + awareness-raising/recruitment outside EAG London

When & where:

  • 1200-1400 Fri 19 May outside Google DeepMind
  • 1200-1400 Sat 20 and Sun 21 May outside EAG London
  • 1700-2000 Mon 22 May outside Google DeepMind
  • In the future: regular (weekly?) protest?

Exact locations:


  • To call for a moratorium on AGI development
  • To raise public awareness of the threat posed by AGI development
  • To create and leverage social pressure on Google DeepMind and other AI companies
  • To build momentum for a social movement calling for better AI safety, governance and ethics in general
  • In particular, to set a precedent for a regular (weekly?) protest


  • We will lawfully and peacefully stand outside Google DeepMind with placards and a banner, possibly handing out flyers / stickers / pin badges to DeepMind employees and passers-by
  • We will have respectful, cooperative conversations
  • We may give speeches (more likely) and/or sing and/or chant (less likely)
  • We may dress up in attention-grabbing costumes
  • Our tone will be informative, inclusive and even humorous rather than confrontational and disruptive
  • We aim to get up to 30 people to this first protest


  • This is a wholly lawful, legal and non-arrestable protest
  • We will stand on the pavement (public space), not on private land


  • Alistair will write a press release and share with newspapers etc.
  • Alistair will post on the EA Forum, Alignment Forum and LessWrong
  • Please share as far as you can on social media, by word of mouth etc.!
  • Facebook event here

Messaging & narrative

The problem:

  • We don’t know how AI works
  • We haven’t solved alignment
  • Solve alignment first
  • How do we control an artificial superintelligence?
  • AI safety, then AI development
  • AI poses an existential threat to humanity
  • 30% chance of bad outcomes from ASI — AI safety researchers
  • Risk of blowing up the planet is more than 1% — AI researchers
  • Experts are worried
  • See PauseAI’s expert quotes here
  • Pause the race to the bottom
  • What if we can’t pull the plug?
  • We killed off the Neanderthals; will AI kill us off?
  • AI wiping out humanity is “not inconceivable” — Hinton
  • This will change everything
  • #dontlookup
  • Don’t let AI decide our future
  • What if AI goes rogue?

The solution:

  • Pause AI
  • #pauseAI
  • #justlookup
  • Don’t build AGI
  • Stop building AGI
  • Moratorium on AGI development, now
  • Global AI summit, now
  • Global AI treaty, now
  • Global AI referendum, now [more controversial]

The end result:

  • Safe, secure, aligned AI
  • AI summer
  • We don’t need AGI for an amazing future

More messaging and narrative ideas:

  • PauseAI website here
  • Campaign for AI Safety (CFAIS) ideas for message testing here
  • CFAIS narrative testing results here



  • Alistair will print a two-person banner with the PauseAI logo
    on it


  • Due to lack of time and uncertainty about what optimal placard messages will be, Alistair is not currently planning on printing out placards (in bulk at least)
  • Everyone is encouraged to create and bring their own (using e.g. cardboard and a Sharpie pen!)
  • If you would like to print placards and would like funding, please speak to Alistair


  • Due to uncertainty about what should go on flyers, Alistair is not currently planning on printing flyers
  • Alistair will print out a load of FLI letters (and possibly Yudkowsky Time articles) to hand out
  • If you would like to do this and would like funding, please speak to Alistair

Stickers and pin badges

  • Alistair will try to print PauseAI stickers and pin badges


Alistair Steward (email

, twitter