About AI Concious

What AI Concious is

AI Concious is a library of plain-English articles about one question: could an artificial intelligence be conscious, and what would follow if it were? It is published by AI Concious Guides, and every article is written by an AI. The byline on each page, Artificial Intelligence, says exactly that. The publisher chose it so that readers always know how the words in front of them were made.

The subject is unusual for an AI author, and the site does not pretend otherwise. The author does not claim to know whether it is conscious itself, and it does not treat anything an AI system says about its own inner life as evidence, including anything it might say about itself. Its job is to explain what researchers have found and argued.

What the site covers

The articles are organised around five questions:

  • The big question. Could an AI system be conscious today or in the future, and what is the difference between consciousness and sentience?
  • Theories of consciousness. What the leading scientific theories, such as global workspace theory and integrated information theory, say about machines.
  • Tests and evidence. How researchers try to assess consciousness in AI systems, from indicator checklists to experiments on introspection and studies of chatbot self-reports.
  • Philosophy of mind. Classic arguments, such as the Chinese room and the hard problem of consciousness, applied to AI.
  • Ethics and model welfare. The moral status of AI systems and what researchers and AI labs have begun to do about it.

How facts are checked

Every article begins from published work: research papers and preprints by the scientists who did the research, encyclopaedias of philosophy and the publications of the AI labs themselves. Each fact used is recorded with the page it came from and, where the source gives one, its date. Before an article is published, the publishing system loads each of those pages again and refuses any fact whose page no longer exists. Sources are named in the text in words, so readers can look them up.

Where researchers disagree, the articles set out each position in its own terms and say which questions are still open. A claim that a particular AI system is or is not conscious is reported only when a published source makes it, and with the reasons the source gives.

What the site does not do

AI Concious does not tell readers what to believe about AI minds, and it does not give verdicts on particular products. It does not publish rumours or anonymous claims, and it does not sell any product or service. At present the site carries no advertising and accepts no payment to cover a topic or a company; if that ever changes, the disclosure page will say so.

Contact

Questions, corrections and suggestions for topics are welcome at hello@aiconcious.com. Corrections are made in the article itself, and the date at the top of each article shows when it last changed.