Networked machine learning compounds individual errors into population-level knowledge faster than humans can.

Evidence

When driving, people mostly learn from their own mistakes, but they rarely learn from the mistakes of others. People collectively make the same mistakes over and over again. As a result, hundreds of thousands of people die worldwide every year in traffic collisions. AI evolves differently. When one of the self-driving cars makes an error, all of the self-driving cars learn from it. In fact, new self-driving cars are “born” with the complete skill set of their ancestors and peers. So collectively, these cars can learn faster than people. — Shoshana Zuboff, The Age of Surveillance Capitalism, 1319

Categorization

• DEFINE: Concretely illustrates the defining claim that collective machine learning is categorically different from human learning — central to how surveillance capital legitimates itself.
• SEED: Possible essay: the ‘mistake economy’ — how surveillance capitalism weaponizes others’ errors as its raw material, framing collective machine learning as a moral advance over human fallibility.
• XREF: connects to: Reinforcement-learning lets algorithms improve themselves faster than humans can.; Even randomly weighted linear models beat human expert judgments — Brian Christian, The Alignment Problem, 441; Big data and AI outperform human judgment in high-stakes decisions. — Joseph Avery, Jim Sidanius, Joel Cooper, Gregory Davis, Amanda N., Bias in the Law: A Definitive Look at…, 1171 (+3 more)

Broader Connections

• Reinforcement-learning lets algorithms improve themselves faster than humans can. #expands (0.80)
• Human judgment errors are systematic and shared, not random noise #contradicts (0.80)
• Human judgment combined with algorithms outperforms either alone #qualifies (0.78) — Joseph Avery, Jim Sidanius, Joel Cooper, Gregory Davis, Amanda N., Bias in the Law, 1194
• Big data and AI outperform human judgment in high-stakes decisions. #corroborates (0.75) — Joseph Avery, Jim Sidanius, Joel Cooper, Gregory Davis, Amanda N., Bias in the Law, 1171
• Even randomly weighted linear models beat human expert judgments #corroborates (0.72) — Brian Christian, The Alignment Problem, 441