Dialog Club's AI Grades Members by Wealth and Fame in Secret

There is something clarifying about a system that makes its social sorting explicit. Dialog, the private invite-only network co-founded by Peter Thiel and entrepreneur Auren Hoffman in 2006, has been doing algorithmically what elite social networks have always done informally: ranking people. Wired reports that the outlet obtained a trove of leaked internal data revealing that Dialog grades its nearly 200 event attendees on a hidden scale, with dossiers compiled by staff and artificial intelligence systems evaluating each member's wealth, fame, influence, and political fit.

The grading scheme runs in a counterintuitive direction: C ratings go to the most prominent figures, A to those who are established but less high-profile, and B to most others. Those grades do real work. They determine in part what attendees are charged to participate in Dialog events, fees that can reportedly run into the tens of thousands of dollars, and they feed an algorithmic matchmaking system that governs both who you meet and where you sit. The dossiers, per the reporting, also track members' apparent political leanings alongside other personal information.

The economist Tyler Cowen offers the clearest window into how the system reasons and where it breaks. The AI reportedly determined Cowen was 'widely recognized within his field' but denied him a top-tier grade because he was not the leader of an organization that is a household name to the average person. That is a very specific failure mode: a model calibrated to organizational prominence that systematically undervalues individual intellectual influence. Cowen is, by most accounts, one of the more genuinely consequential thinkers in American policy circles, and the AI simply did not have a category for him.

The honest caveat is that Wired's source is a single confidential leak, and Dialog has not had a meaningful public chance to respond to the specifics. What the reporting does not give you is how common this approach is among similar private networks, how many of the nearly 200 people named in the data know their dossiers exist, or whether Dialog obtained any consent under applicable privacy law for AI-generated profiles that reportedly include apparent political leanings and other personal information.

For practitioners thinking about AI in social and professional contexts, the Dialog case is useful not because it is monstrous but because it is banal. This is a predictable application of AI to a gatekeeping problem that humans were already solving with spreadsheets and gut instinct. The AI did not introduce the ranking; it industrialized it and made it legible. That legibility is, paradoxically, what created the exposure. Privacy regulators looking for a clean case study in AI-generated personal dossiers and social scoring now have one on the record.