AI models agreed with each other more than with reality
A new test from Science Prediction Universe found that pooling AI rankings improved average results, but shared scientific errors survived the crowd.
Advertising disclosure: we may earn a commission when you join an operator via links on this page, at no cost to you. This never dictates our ratings. How we make money
Key takeaways
- Science Prediction Universe tested AI models on scientific prediction.
- Pooling the models' rankings improved average results.
- Errors the models shared didn't go away when their answers were combined.
- Agreement between models isn't proof that an answer is correct.
Stacking AI models helps. Only so far. A new test from Science Prediction Universe found that pooling the models' rankings improved average results, but the scientific errors they shared made it through untouched.
What the test found
Start with the good news. Group the rankings together and the average gets better. That's the standard crowd effect, and it showed up here right on cue.
Now the catch. When several models made the same mistake, pooling didn't scrub it out. The models agreed with each other more than they agreed with reality. Awkward.
The source material is thin. It doesn't say which models were tested, how many, or how big the gains were. Treat the details as unconfirmed until the full study lands.
Why it matters
Crypto leans harder on AI every month: research summaries, trading signals, autonomous agents that act on their own conclusions. The popular fix for flaky output is to ask several models and take the consensus. This test says that fix has a ceiling.
If models learn from similar data, they can be wrong in the same direction. Consensus then looks like confirmation when it's really correlated error. If you hold, trade or build on AI-generated analysis, that's your risk. Agreement feels like safety. It may not be.
So check AI claims against real data, not just against another AI. And ask yourself how much independence you're really getting when most of these models learned from the same internet.
Source: CryptoSlate
Top Crypto Casinos Right Now
🇦🇺 Showing sites that accept players from Australia Change country
Advertising disclosure: we may earn a commission when you join an operator via links on this page, at no cost to you. This never dictates our ratings. How we make money
More info Less info
Why we chose it: A US-oriented casino that turns away only New Jersey. The cashier takes crypto, but the site publishes fewer specifics than any operator we list.
- KYC
- Required
Restricted countries: New Jersey (US).
More info Less info
Why we chose it: One of the widest crypto line-ups of any book we list - seven coins with fast payouts - aimed squarely at Asia and Latin America rather than the US or Western Europe.
- Licence
- Curacao
Restricted countries: Austria, Belgium, Bulgaria, Cyprus, Czechia, Germany, Denmark, Estonia and 21 more.
More info Less info
Why we chose it: Built for American bettors, priced accordingly. Biggest welcome bonus of the five, attached to the heaviest rollover of the five.
- Online since
- 1994
- Licence
- Curaçao
- Payout time
- 24–48 hours (crypto)
- Wagering
- 18x sportsbook / 30x casino
- KYC
- Required
Restricted countries: Belgium, Costa Rica, Croatia, Curaçao, Aruba, Sint Maarten, Caribbean Netherlands, France and 11 more.
Alex Rivera
Alex covers crypto markets and trading, testing every platform hands-on before writing about it.
Related News
Maternal AI chatbot 'Rosie' fails to beat a kids' book program
Nonprofits sue OpenAI, Anthropic, Google to freeze AI development
ACE Music v1 Tops AI Music Rankings, Edging Out Suno in Decrypt Test