TIME100 Innovators share the thing they wish more people understood about AI.
Category
🗞
NewsTranscript
00:00 (upbeat music)
00:02 AI is often described as something alien.
00:14 However, I prefer to think about it as us in aggregate.
00:17 It's a monumental collective accomplishment
00:20 trained on all of the human knowledge
00:21 that's archived online.
00:23 Humans should get more credit and we're working on that.
00:26 - One thing that I love to the broader public
00:28 to understand better about AI
00:30 is that AI, like many other forms of technology,
00:34 is a tool for humanity.
00:37 It's not inherently good or bad
00:39 in the same way that a shovel or a pickaxe isn't good or bad,
00:44 but it's about the way we use the technology
00:47 and make sure that it's used for the right purposes
00:51 in order to promote and push forward progress.
00:54 That is where the consequences
00:57 and potentially a lot of the upside comes as well.
01:00 - AI will impact almost every job out there
01:03 in the next five to 10 years,
01:05 but at the same time, that doesn't mean AI is conscious,
01:09 has a mind of its own,
01:10 is an existential threat to humanity,
01:13 because no one is working on AI doing whatever it wants.
01:17 Mostly it's about people and the systems
01:20 and the frameworks and the data that they use in AI
01:23 and that AI is part of that we have to worry about
01:27 and that's the AI itself.
01:28 - Instead of thinking that AI is some abstract thing
01:30 in the sky or in the cloud
01:32 that operates without any human intervention,
01:36 especially when we think about the goals of AI systems,
01:39 they're all specified by humans,
01:40 and so our ability to control or reduce some of the concerns
01:45 that folks have around AI are really achievable
01:49 by working on what a lot of people in the field
01:51 are calling alignment.
01:52 - I think that AI is often portrayed as a black box
01:55 and I wish people understood
01:57 that we can actually open up this box
01:59 and get a clear picture of the decision-making processes
02:03 in AI models.
02:04 It's not magic, really.
02:06 It's just well-crafted algorithms
02:08 that are used to explain patterns in large scales of data
02:12 that humans just can't always manually identify.