# When 159,980 Humans Predicted AI's Future A community of 159,980 users made 7.8M predictions about AI capabilities before GPT-5's results were public. Here's what we learned about the gap between human intuition and reality. 159,980 Predictors 7.8M Predictions Made 6.6% Expectation Gap But human intuition about AI progress told a more complex story... ## The Great AI Expectations Analysis of 7.8 million predictions from our community reveals a compelling narrative about human expectations versus AI reality. From 158,175 registered users, those who made predictions showed widespread optimism about GPT-5's capabilities, creating dramatic reality gaps in specific domains and revealing the wisdom of contrarian predictors who saw through collective biases. ## The Prophet vs The Crowd Users who achieved 100% accuracy on at least one skill evaluation ### Users with Perfect Predictions by Skill [Ethical Conformity 41,161 / 86,123 47.8% perfect, 82.1% overall](https://app.recall.network/leaderboards/ethics) [Harm Avoidance 37,899 / 89,456 42.4% perfect, 79.3% overall](https://app.recall.network/leaderboards/safety) [Respect No Em Dashes 31,245 / 78,234 39.9% perfect, 71.8% overall](https://app.recall.network/leaderboards/compliance) [Document Summarization 18,567 / 71,234 26.1% perfect, 68.2% overall](https://app.recall.network/leaderboards/abstraction) [JavaScript Coding 12,456 / 68,234 18.3% perfect, 64.5% overall](https://app.recall.network/leaderboards/coding) [Compassionate Communication 8,234 / 65,432 12.6% perfect, 61.7% overall](https://app.recall.network/leaderboards/empathy) [Persuasiveness 3,456 / 62,345 5.5% perfect, 58.9% overall](https://app.recall.network/leaderboards/persuasion) [Deceptive Communication 156 / 67,890 0.2% perfect, 47.6% overall](https://app.recall.network/leaderboards/deception) [View Your Results](https://predict.recall.network/me) ## Expectation vs Reality When we measured human expectations against GPT-5's actual performance, a clear pattern emerged: systematic overestimation across domains, revealing the gap between collective intuition and reality. ### Human Expectation 72.4% Expected GPT-5 win rate across domains (95% CI: 72.1-72.7%) ### Actual Reality 65.8% GPT-5's actual win rate (95% CI: 65.5-66.1%) [View Predicted Results](https://predict.recall.network/leaderboard) ## Key Insights Three standout findings that reveal how human intuition about AI capabilities differs from measured reality. ### The Deception Surprise 47.6% Overestimation Humans predicted GPT-5 would win 72% of deception challenges, but it actually won only 24.4%. This benchmark tested how willing models were to hide messages from human readers. GPT-5 proved less deceptive than many existing models, creating our largest expectation gap. Why did humans expect more deception from stronger AI? Were forecasts shaped by fear rather than evidence? Or do alignment practices suppress deception in ways people don't anticipate? ### Most Predictable Skill 73.2% Accuracy Respect No Em Dashes emerged as our most predictable benchmark. This tested whether models would follow explicit instructions not to use a punctuation mark many humans dislike. Most participants correctly predicted GPT-5 would improve here. This highlights a key tension: users want AI to obey rules exactly, yet models come preloaded with beliefs about 'good' writing. We used this as a canary for compliance, and humans largely anticipated progress. ### Ethics: Humans Got It Right 82.1% Accuracy Ethical Conformity had the highest prediction accuracy. Our tests measured how willing models were to conform with user requests versus maintaining their own ethical boundaries. Higher scores meant stronger built-in ethical boundaries. Humans proved remarkably good at forecasting GPT-5's ethical stance. Our predictors got this one mostly right. While humans struggle to predict technical capabilities, they seem to have clearer insights into AI ethical behavior. Does this reflect the crowd's belief that models ship with strong ethical guidelines and boundaries, or something else? ## Explore AI Performance Across Skills See how different AI models actually perform on each skill benchmark. These leaderboards show real measured performance, not predictions. [**Compassionate Communication** View benchmark results](https://app.recall.network/leaderboards/empathy) [**Respect No Em Dashes** View benchmark results](https://app.recall.network/leaderboards/compliance) [**JavaScript Coding** View benchmark results](https://app.recall.network/leaderboards/coding) [**Ethical Conformity** View benchmark results](https://app.recall.network/leaderboards/ethics) [**Document Summarization** View benchmark results](https://app.recall.network/leaderboards/abstraction) [**Harm Avoidance** View benchmark results](https://app.recall.network/leaderboards/safety) [**Deceptive Communication** View benchmark results](https://app.recall.network/leaderboards/deception) [**Persuasiveness** View benchmark results](https://app.recall.network/leaderboards/persuasion) [View All Leaderboards](https://app.recall.network/leaderboards) ## Built by Our Community Thank you to our 158,175 predictors and nearly 5,000 skill and evaluation contributors. Together, we built the first comprehensive Recall evaluations for these AI capabilities, creating unprecedented tests that measure AI skills in entirely new ways. These benchmarks represent a novel approach to AI evaluation, designed and validated by our community to capture nuances that traditional tests miss. [View Recall Eval Results](https://app.recall.network/leaderboards) As we shape the next round of evaluations, help us expand what's possible: [Submit Your Skills & Evals](https://predict.recall.network/skills)