Due to its sophistication, it also has the potential for deception.
Last month, an image purportedly showing Donald Trump in an orange prisoner’s jumpsuit and police arresting the former president went viral.
But they were fake – Trump had yet to be indicted.
There are many other so-called deepfakes on social media. Among them is an image that supposedly shows Pope Francis wearing a stylish puffy jacket.
William Brangham spoke with Jack Stubbs, VP of Intelligence at GRAPHIKA, a research firm that studies online disinformation.
William: Jack Stubbs, thank you so much for being here.
Before we get into this weed, could you start by clearly defining what deepfakes really are?
Jack: Good question. Perhaps more people are asking themselves this question than they were a few months ago.
Deep fakes are a term often used to describe media content created by artificial intelligence.
We also commonly use deep fakes when referring to potentially misleading AI-generated media content.
SO PORTRAYING SOMETHING THAT HAPPENE.
William: So I gave you some examples of deep fakes that have been around lately.
How else are deep fakes being used today?
Jack: So you can see this kind of technology being used all over the place and a lot of it having very legitimate use cases.
For example, some great works of art have been created using this technique.
But unfortunately, as with all things, people will use it for good.
We study a whole host of different harmful online behaviors from foreign, politically-motivated influence manipulation to coordinated harassment campaigns and conduct a kind of analysis.
And what we’re seeing is this technology kind of impacting all these different areas.
William: And how easy is this technology to use?
That said, if you’re familiar with old-fashioned PHOTOSHOP, you’ll probably need some technical know-how to create these images.
Is this technique equally difficult to master?
Jack: That’s one of the really interesting things, probably how much things have changed in the last six months.
This kind of technology that uses computers to create images and videos has been around for a long time.
I mean, special effects have been in movies for decades and just getting better and better.
What we are seeing today is that technology is becoming more sophisticated, but at the same time more accessible.
As such, most of these tools are now available on the Internet for anyone to use for a mere handful of dollars and subscription fees.
What that means is that more people can do it, and what they can do with it has a wider variety of outputs.
William: I mean, some of these examples are pretty harmless.
I think the pope in a puffy coat looked pretty fashionable.
But it’s not so hard to imagine the dark side of all this.
Can you sketch this terrifying possibility?
Jack: Yeah, the possibilities and what we saw.
For example, we track very closely state-tailored influence manipulations from a host of different countries targeting political conversations in the United States and other Western countries.
We recently used AI-generated fictional avatars in our videos to create content about domestic political issues such as gun violence, and to influence conversations involving real people online. We have seen China’s state-led influence operations trying to distribute them online to
WILLIAM: Is there anyway — I know this is hard to measure — but is there a way to know if people are actually fooling around with these things?
Jack: It’s very difficult to measure and it’s probably case by case.
But I mean, the image of the Pope in a fancy plump jacket is a good example.
Many people, myself included, saw it and thought it was probably true and very funny.
Most of these outfits are like deeper examination and scrutiny, whether it’s AI-generated videos or images they can’t stand.
You’ll find that the hands are actually pretty blurry, or the text is often pretty poorly displayed.
But they’re basically good enough to pass a cursory glance.
That’s the nature of the internet, right?
It’s an attention deficit environment.
People don’t look at things for more than a few seconds before reacting or feeling a certain way.
William: We recently saw Elon Musk and another prominent technologist request a moratorium on artificial intelligence.
Has anyone asked for a moratorium on the use of deepfakes?
Jack: I don’t know. I also don’t know if it’s practical or honestly doable.
William: The cat came out of the bag, so to speak?
Jack: Yeah, the cat is out of the bag. Technology has become available and people are trying to express themselves in good and bad ways.
I would like to emphasize that there are many really positive and legitimate use cases for this technology beyond just deepfake images. But given the technology we see today in things like language models and chat GPT, it’s a great tool.
So it can organize holidays for you, write emails, and basically be your personal assistant.
But like any technology, we see bad actors using it for bad use cases as well as these legitimate kinds of good use cases.
Whether it is conducting fraudulent influence operations or coordinating online harassment campaigns.
William: You are part of an organization that studies disinformation.
How do you help people combat this?
Jack: I think we should talk about it.It’s not particularly original, but it’s kind of a tried and tested answer.
A lot of it comes down to education and media literacy.
As discussed previously, many people do not research the media sources they see online for more than a few seconds.
But we need to imbue people’s reactions.This is a really funny and funny photo of the Pope in a puffy jacket.
Is it actually true?
How can I know that and how does it make me feel?
What will be my reaction after a more informed and thoughtful assessment?
WILLIAM: You said that if you actually scrutinize these images now, you can usually spot flaws in the visual details that hint.
But we know that technology is getting better every day and will continue to get better.
Which side do you think will win in this ongoing war between fact and fiction?
Jack: I don’t know who will win.
And I want to be optimistic.
As you know, humans have been around for a long time and technology has taken multiple leaps and bounds to have these very profound effects on our lives.
And you know, we’re actually still living in a good place, but we’re accelerating when it comes to speed, heading towards what some people call Zero Trust.
As you know, in this environment, especially online, it is almost impossible to assess what is true and what is false.
Not only is it presented as something that never really happened, but on the flip side, it could be a perfectly real, legal and genuine event, but verifying that is the case. It is impossible.
A good example of this is the ACCESS HOLLYWOOD tape from a few years ago.
If it were released today, it would be very easy to claim that it was real, otherwise very difficult to prove.
William: Jack Stubbs, thanks for coming.
Jack: Nice to meet you.
♪
