Staging draft for review. Not live, and the backers shown are placeholders. Please don't share publicly.

Fair Future Safeguards · Ask 3 of 5

A human is always available.

You're told when you're dealing with AI, and a person can always step in.

What happened

“The system decided.”

More and more of the decisions that shape a life are made by software. Whether your benefit claim gets flagged. Whether your CV gets past the first sift. Whether your loan is approved. How long you wait for an appointment.

Often nobody can tell you why. Often you don't even know a machine was involved. Ring up to ask and you may find the phone line has been replaced by a chatbot that can't help and can't put you through to anyone who can.

“The system decided” is not an answer anyone should have to accept.

What we're asking for

Told, explained, reviewed.

The ask, in full. You are always told when you're dealing with AI. Any automated decision that affects you must be explained, open to challenge and reviewable by a person. Every public service must offer a human.

What changes for you.

  • Every AI you deal with says so. No more guessing whether you're talking to a person.
  • If a machine makes a decision about your money, health, work or credit, you get the reasons in plain English.
  • You can challenge it, and a real person with the power to overturn it has to look.
  • Every public service keeps a human route: a phone number, a counter, a person. Automation can speed things up, as long as it's never the only door.

Who already does this.

UK data law already gives you some rights over purely automated decisions, but they're narrow and hard to use. The EU's AI Act requires companies to tell you when you're interacting with AI, and gives people a right to an explanation of decisions made by high-risk systems. We'd make those rights simple and universal.

Why this is good for AI.

This is the safeguard that lets public services actually use AI. Councils and hospitals are cautious because one bad automated decision with no way to appeal is a scandal. Give people a clear right to an explanation and a human, and the boring parts can be automated with a clear conscience. The robot can do the paperwork. It can't be the appeals process.

Yes to AI, on fair terms.

Never argue with a machine that can't hear you. Add your name.

Sign the petition.

The five asks

The other four.

Ask 1

Your data is yours.

No company trains AI on your posts, photos or messages unless you actively say yes.

Read this ask.
Ask 2

Your face and voice are yours.

A legal right to stop anyone making an AI copy of you without your permission.

Read this ask.
Ask 3

A human is always available.

You're told when you're dealing with AI, and a person can always step in.

Read this ask.
Ask 4

Clean power, paid for by them.

Data centres that take public support run on clean energy and never add to your bills.

Read this ask.
Ask 5

Free AI for everyone.

A public AI helper with no adverts, and the skills to use it taught in every school and library.

Read this ask.
All five

Back all of them at once.

One signature covers the lot, and you decide whether we're allowed to email you.

Sign the petition.