https://www.youtube.com/watch?v=Ey0q5CmIV1E&ab_channel=NorgesBankInvestmentManagement

Shantanu 对AI的看法：

What I always say is people who use AI will replace people who don't use AI. Everybody talks about how this is going to replace humans. Well, the place where AI is probably the most advanced is in software writing.

When you think about what's happened with the ability to say, "I want to create a program," that's what's the most expensive. I don't know if at your institution, you're reducing the number of software engineers and developers. We're all saying, how do we make them more productive? How do we allow them to do more at the same cost? So there will be an efficiency increase.

Like every generation of technology, some things will get disrupted, but overall it creates more opportunity. And I certainly believe that for creativity.

对deepfakes的看法：

N: How worried are you about Deepfakes? You should be the world champion in Deepfakes.

That's a great question and I'll get to that. But the thing I wanted to say on that previous question also was that you need a soul for creativity. And I think that's the part where human ingenuity will always be good.

Deepfakes, we actually started when we thought about Adobe and we said, listen, we create the world's content and it's really critical for us to also take responsibility for what's happening with content. And so we created this thing, it's called content credentials. Think of it as a nutrition label, Nikolai. And when we started this, we said, we have one part in the chain, but we have now got thousands of partners who've signed up for this content authenticity initiative as it's called.

And what the content authenticity initiative is, camera manufacturers, when they take a picture, they're putting a digital signature on that picture saying it was created with this camera at this time taken by Nikolai. When you edit it in a piece of software, we stamp that with what's happened. If somebody who's in the content distribution game, whether it's Meta, whether it's New York Times, whether it's the BBC, they've all signed up to carry those credentials. So I think the plumbing and the infrastructure for Deepfakes is all there.

I'll tell you what the real challenge is. So when you talk about Deepfakes, the ability to look at the provenance of a piece of content is all there. The question is, will a consumer want to see where that came from, who created it, who might have changed it? And so now we have this ability, we're working with the U.S. government, 50%. I think it's 50% of the world's population is going to have an election, I think, in a 12-month period. Some of them have certainly happened.

And so in that particular context, what we are saying is when you see a piece of content, you can see this label, it says content credential. Think of it as a nutrition label. If you click on it, it will tell you who created it, when it was created, when it was perhaps altered. The question is, how many people actually read the nutrition label when you're eating food? And it's the same issue with content. How many people are going to say, I want to know where that piece of content was created?

So I would say the technology companies, the people distributing this media, we've all participated in the Deepfakes lifecycle to understand how we can bring that value. Now the consumer has to say, when somebody that you know is saying, I want to see that piece of content, are they going to insist on evaluating the content credential? Because neither you nor I can come in the way of people creating content. It's the question is the consumer, how much do they care about where they came from? Hopefully that makes sense.