A recent presentation at Fordham’s cybersecurity conference last month helped illustrate why.

“When I asked the CEO of a major movie company recently, ‘What’s the craziest thing you can imagine will happen in the next two to three years?’ he said, ‘We will have a full cinematic feature starring zero actors, zero cinematography, zero lighting, and zero set design,” said Josh Wolfe, co-founder and managing director of Lux Capital at a keynote speech on Jan. 10.

“It will all be generated.”

As an example, Wolfe, whose firm invests in new technologies, screened a fan-made movie trailer that used AI to imagine what Star Wars would look like if it had been directed by Wes Anderson.

A Threat to Storytelling

James Jennewein, a senior lecturer in Fordham’s Department of Communication and Media Studies whose film-producing credits include Major League II, Getting Even with Dad, and Stay Tuned, said the prospect of AI-powered screenwriting is deeply concerning.

He called storytelling “soul nourishment” that teaches us what it means to be human.

“We’re still watching films and reading books from people who died centuries ago, and there’s something magical about an artist digging into their soul to find some kind of truth or find a unique way to express an old truth, to represent it to the culture, and I don’t think that AI is going to help make that happen more,” he said.

In many ways, AI has already infiltrated movies and TV; major crowd scenes in the show Ted Lasso were created using AI tools, for example. This summer, the directors of Indiana Jones and the Dial of Destiny used AI to render the nearly 80-year-old Harrison Ford to look like he was in his 20s.

The ability to use fewer actors in a crowd scene is obviously concerning to actors, but Jennewein said the strike was about more than just saving jobs–it’s about protecting creativity.

“We don’t want AI to create the illusion that something is original when it really is just a mashup of things that have been created before,” he said.

“Flesh-and-Blood” Films Coexisting with AI

Paul Levinson, Ph.D., a professor of communications, saw first-hand what AI can do to his own image and voice. A 2010 interview he did was recently altered by the journalist who conducted it to appear as if Levinson was speaking in Hindi.  But he is less concerned about AI taking over the industry.

He noted that when The Birth of a Nation was first screened in 1915, it was predicted that it would kill off the live theater.

Levinson predicted that in the future, the majority of what we watch will be AI-generated, but there will still be films that are made with live human actors. Just as theater co-exists with live movies, traditional movies will co-exist with AI content.

“I think we are going eventually to evolve into a situation where people aren’t going to care that much about whether or not it’s an AI-generated image or a real person,” he said.

Levinson acknowledged that AI could inflict real harm on the livelihood of actors and screenwriters, but said an equally important concern is whether those who work with AI tools get the credit they deserve.

“I’m sure people are going to think I’m out of my mind, but I don’t see a difference, ultimately, between a director who is directing actors in person and somebody who understands a sophisticated AI program well enough to be able to put together a feature-length movie,” he said.

“What could ultimately happen as AI-made films become more popular, is that films that are made with real flesh-and-blood actors will advertise themselves as such, and they’ll try to do things that maybe AI can’t quite yet do, just to push the envelope.”

And misinformation, they said, has the potential to do enormous damage.

“It’s a threat because what you’re trying to do is educate the citizenry about who would make the best leader for the future,” said Karen Greenberg, head of Fordham’s Center on National Security.

Greenberg, the author of Subtle Tools: The Dismantling of American Democracy from the War on Terror to Donald Trump (Princeton University Press, 2021), is currently co-editing the book Our Nation at Risk: Election Integrity as a National Security Issue, which will be published in July by NYU Press.

“You do want citizens to think there is a way to know what is real, and that’s the thing I think we’re struggling with,” she said.

At the International Conference on Cyber Security held at Fordham earlier this month, FBI Director Chris Wray and NSA Director General Paul Nakasone spoke about the possibility of misinformation leading to the chaos around the U.S. election in a fireside chat with NPR’s Mary Louise Kelly. But politics was also a theme in other ICCS sessions.

Anthony Ferrante, FCRH ‘01, GSAS ‘04, global head of cybersecurity for the management consulting firm FTI, predicted this year would be like no other, in part because of how easy artificial intelligence makes it to create false–but realistic—audio, video, and images, sometimes known as deepfakes.

The Deepfake Defense

“I think we should buckle up. I think we’re only seeing the tip of the iceberg, and that AI is going to change everything we do,” Ferrante said.

In another session, John Miller, chief law enforcement and intelligence analyst for CNN, said major news outlets are acutely aware of the danger of sharing deepfakes with viewers.

“We spend a lot of time on CNN getting some piece of dynamite with a fuse burning on it that’s really hot news, and we say, ‘Before we go with this, we really have to vet our way backward and make sure this is real,’” he said.

He noted that if former President Donald Trump were caught on tape bragging about sexually assaulting women, as he was in 2016, he would probably respond differently today.

“Rather than try to defend that statement as locker room talk, he would have simply said, ‘That’s the craziest thing anybody ever said; that’s a deepfake,” he said.

In fact, this month, political operative Roger Stone claimed this very defense when it was revealed that the F.B.I. is investigating remarks he made calling for the deaths of two Democratic lawmakers. And on Monday, it was reported that days before they would vote in their presidential primary elections, voters in New Hampshire received robocall messages in a voice that was most likely artificially generated to impersonate President Biden’s, urging them not to vote in the election.

A Reason for Hope

In spite of this, Greenberg is optimistic that forensic tools will continue to be developed that can weed out fakes, and that they contribute to people’s trust in their news sources.

“We have a lot of incredibly sophisticated people in the United States and elsewhere who understand the risks and know how to work together, and the ways in which the public sector and private sector have been able to share best practices give me hope,” she said.

“I’m hopeful we’re moving toward a conversation in which we can understand the threat and appreciate the ways in which we are protected.”