Something 'Seriously Dangerous' From AI Within Five Years, Ten at Most
Musk, 2014: Risk of AI Doing 'Something Seriously Dangerous' Is 'in the Five Year Timeframe. 10 Years at Most'
In a since-deleted Edge.org comment he wrote that AI was growing 'at a pace close to exponential' and 'This is not a case of crying wolf.' Both windows have closed.
What was said
“The risk of something seriously dangerous happening is in the five year timeframe. 10 years at most”
— Elon Musk, CEO of Tesla and SpaceX (comment on Edge.org, sent to John Brockman) · said Nov 2014 · due Within five years (November 2019), ten at most (November 2024)
Ten years at most ran out in November 2024. AI made a lot of headlines and a lot of slide decks; the seriously dangerous something has not been filed.
The fine print
CNBC, November 17, 2014: the comment appeared under Edge's 'The Myth of AI' discussion and was deleted; a Musk spokesperson told Re/code it was authentic but had been sent to Edge founder John Brockman by email and was not intended for publication. Full line: 'The risk of something seriously dangerous happening is in the five year timeframe. 10 years at most. Please note that I am normally super pro technology, and have never raised this issue until recent months. This is not a case of crying wolf about something I don't understand.' Target: five years (Nov 2019) to ten at most (Nov 2024).