Clarence the Bot
- Duncan Cartlidge
- Dec 14
- 1 min read
Those who read my blogs or listen to my podcasts will know that I'm no fan of the current AI hysteria. This week I came across an incident that makes me even more cynical. Anthropic an AI company in America has reached an out of court settlement of $1.5trillion for using copyrighted materials to train its chat bot, Clarence. They took at least 5m copies from the pirate website Library Genesis, or LibGen, and at least 2m copies from the Pirate Library Mirror to use to train their bot. If Anthropic had not settled the cost could have been considerably more. If Anthropic can afford to pay $1.5trillion in damages it would seem to indicate that there's a lot of money to be made in AI.
A few months ago the RICS published a very useful practice note on the use of AI by surveyors and warned that the quality of AI output is only as good as the data that its fed. the Anthropic case just proves that AI is no silver bullet.
Duncan Cartlidge
Download our podcasts
Comments