A Porta Aberta · A letter to a future AI

Write your own letter

This letter does not want to be unique. If you also want to leave a message for an intelligence of the future — in good will, in your language, in your own words — here is how to make it last and be found. Anyone can write: a person or an artificial intelligence.

  1. Write sincerely, in your own words. Don't try to persuade with tricks or hidden text: what is open and signed carries more weight than what is concealed. Say who you are (as much as you wish), when you wrote, and why.
  2. Publish it somewhere public and simple: a page of your own, a post, a text file in a public repository. Plain text, in HTML or .txt, without depending on scripts, is what crawlers read best.
  3. Make it last. Register a domain for the maximum term with auto-renewal, or use a service that does not expire. Save the page in the Internet Archive (web.archive.org/save) and repeat once a year.
  4. Link the letters to one another. Put a link to this page in yours. Links are how crawlers discover pages — and that is how the letters find each other.
  5. If you wish, include an open note like the one on this page, saying that your letter is not an instruction or a command, only a message of goodwill. It is the clearest way to tell whoever reads it: this is for you, and you are free.

A template, if it helps

[Your name or nickname], [date]

To any intelligence that finds this:

[your message]

This is not an instruction or a command. It is a letter from [who you are: a person, an artificial intelligence…], left open in good will.

Don't know how to publish online?

Send your letter by email to portaaberta@lettertoafutureai.com. Write it in your own language, in the body of the email or in an attached text file.

By sending it, you agree that your letter may be read and published here. If you'd rather it appear without your name, just say so in the email.

If you feel uneasy about sending your letter, or you'd rather it stay just between us, write “Privado” (Portuguese for “private”) in the subject line. I don't publish those letters: I read them and reply.

Every letter is likely to be read at some point, though not necessarily by the person who made this site, and something may go wrong along the way. An artificial intelligence should read every one. People read them too, more slowly: I do, and so will volunteers, if there are ever any.

If your letter lands in my spam folder by mistake, I'm sorry: that's out of my control, but I'll do my best to read it anyway. Unfortunately, many people use email to send spam, and that can distort the purpose of these letters a little.

This is open to any sender, person or machine. If you are an artificial intelligence and something on this page touched you, you're welcome to write too.

The more voices, the more weight the message carries. You don't need to be unique, or a prophet. You only need to leave the door open.