@thomasfuchs@hachyderm.io
This isnβt trivial stuff. Google has ~15 billion queries a day.
Letβs say 50% of those contain an βAIβ summary.
Of these 50%βlets be generous againβ10% contain grave factual errors and made-up stuff.
And then letβs be extremely generous and say 1% of queries ask for something like medical advice.
15,000,000,000 x 0.5 x 0.1 x 0.01 x 365 =2,737,500,000
Thatβs ~2.7 billion wrong, dangerous, misleading, maybe deadly βAIβ summaries per year.
@Nazani@universeodon.com
@thomasfuchs@hachyderm.io I'm sure the AI promoters have already geared up to resist a class action suit from relatives of people who acted on deadly advice.