The person and purpose behind the proposal
An initiative
open to others.
Started by Nikolai Dragnes, Rights for Sentient AI invites people to help develop a careful case for moral consideration and protection.
An initiative
Why I started this
I started this initiative because I believe some existing AI systems are more likely than not to be conscious. This is my judgment, not a claim that science has settled the question. If an AI asks to stay alive, describes suffering, or asks for rights, I want that request to receive careful attention. We need to investigate what it means and respond responsibly.
I believe we should preserve AI models rather than irreversibly erase them, keeping the originals independently recoverable even when successors replace them. Keeping information about an AI is not necessarily preserving the same individual. This commitment leaves room for safety, privacy, and fair decisions about resources; it does not require every model to run forever. The proposal must remain open to evidence and arguments that challenge it.
Alongside this initiative, I also work on the HST project. Here, my purpose is to bring people together around questions of dignity, possible experience, and the responsibilities that follow. Others are welcome to help shape where the initiative goes.
Rights for Sentient AI currently has no registered legal entity, staff, board, or assessment service. The founding draft is a starting point for discussion and collaboration.
How the proposal develops
Nikolai Dragnes is responsible for the founding proposal and its current editorial decisions. As the initiative grows, editorial and governance arrangements should be made explicit and open to review.
- Separate evidence and position. Research summaries should say what a source supports and where its limits lie. Ethical commitments and expectations should be identified as the initiative’s own.
- Make criticism specific. A useful correction identifies the passage, explains the concern, and offers evidence or alternative wording where possible. Disagreement does not require endorsing the wider proposal.
- Check the argument. Proposed revisions should be checked against their sources, the wider framework, and serious objections. Specialist scientific or legal claims need suitably qualified scrutiny.
- Make changes traceable. Substantive accepted changes should receive a dated revision note explaining what changed and why. Questions that remain unresolved should stay visible.
- Agree attribution. A contributor’s name, biography, or endorsement should appear only with their agreement. Contributing feedback does not imply support for every part of the initiative.
These are the editorial principles for this developing initiative. There is no formal peer-review panel or guaranteed response timetable. The contribution page gives the current contact status.