Not at all. Cars and roads are fantastically dangerous, many thousands of people die every year, but we still build cars and roads.
It's the same for AI. We need to treat the risks responsibly which means researching them and making informed judgements. That's what he's talking about.
> Cars and roads are fantastically dangerous, many thousands of people die every year, but we still build cars and roads.
The statistics aren't that straight forward; for example, young men under the age of 24 are significantly over-represented in traffic deaths, so it's not entirely reasonable to assume the cars or roads are inherently dangerous. On top of that, we drive 3.1 trillion miles every year in the U.S alone and falling off a ladder at work kills about twice as many people than roadway fatalities do.
Worldwide.. and that table is out of date, both catagories have increased in the new table. I can only speak to the statistics in the U.S. where ~36,000 people died according to NHTSA's FARS database. Of those, 6,000 were pedestrians. Whereas ~33,000 people died from falls or related causes according to the CDC. So, my ratio was wrong.. but I don't think it diminishes my point too significantly.
Falls disproportionately affect the elderly.. as do traffic accidents, but the opportunities for risk are typically fewer as many elderly stop driving at some point, most die as passengers when they're involved in traffic accidents.
But self driving cars will kick a lot of drivers out of their jobs. I cannot see how his logic doesn't apply to his own business, maybe that is why he is not emphasizing it?
It's the same for AI. We need to treat the risks responsibly which means researching them and making informed judgements. That's what he's talking about.