Dissensus has been accepted as an affiliate of the International Association for Safe and Ethical AI and is listed on its affiliates page. For a lab this size it is the most substantial institutional step so far, and it points where the work is already heading.
The company is the reason it matters. The affiliate list runs to a few hundred organisations and includes MIRI, METR, FAR.AI, MATS Research, the Cooperative AI Foundation, Harvard’s Berkman Klein Center, Oxford’s Institute for Ethics in AI, Berkeley’s Center for Long-Term Cybersecurity, Cambridge’s Leverhulme Centre for the Future of Intelligence, and MIT FutureTech. These are groups working on the same failure modes from directions this lab does not come at them from.
Now the part that is easy to overstate, so it is better said here than left for someone else to point out. Affiliation means an organisation broadly supports the association’s mission. That is the whole of the requirement. It is not a grant, not a review of anything Dissensus has published, and not a vote on the association’s positions — those are settled by its elected council. IASEAI’s own terms are explicit in both directions: affiliates are free to hold different positions and pursue different tactics, and no affiliate is described as endorsing any activity or publication unless it opts in. Being on a list next to MIRI is not a claim to be doing MIRI’s work.
With that said, the direction is real. The lab’s AI work — emergent misalignment in multi-agent systems, how collective failure arises from interaction rather than from a single misaligned policy, and the harder ethical questions about entities whose behaviour is learned rather than specified — is the part of the programme that grows from here. Affiliation gives that work a room to be argued in, which is worth more to a one-thesis lab than agreement would be.
More on how the lab works with others, including what this affiliation is and is not, on the About page.