IBM‘s question-answering system Watson won a televised Jeopardy! challenge against the show’s two most successful champions, Ken Jennings and Brad Rutter. The three-day contest concluded on 16 February 2011.
What happened
- Watson finished with $77,147, compared with $24,000 for Jennings and $21,600 for Rutter; IBM gave its winnings to charity.
- The system ran IBM’s DeepQA software on 90 servers with 2,880 processor cores.
- It interpreted clues posed in natural language, running hundreds of algorithms in parallel to find candidate answers and rank them by confidence, typically within about three seconds.
- Watson made some notable errors, including answering ‘Toronto’ to a clue about US cities.
Why it mattered
The win was a high-profile public demonstration of AI understanding everyday language, watched by millions of viewers. IBM went on to market Watson for healthcare and business uses, with mixed results.
Lessons for organisations
When evaluating AI tools, look beyond headline demonstrations to accuracy on your own data, error handling, and how confidence is communicated. Human review remains important where mistakes carry real consequences, and staff should understand the tool’s limits before relying on it.
Part of our Top stories archive of headline-making events in information security, privacy, and AI. If you would like help applying the lessons to your organisation, contact us.
