Friday, 2 September 2022

When Your Kubrick AI Isn’t HAL… It’s an AI Kubrick

NAB

AI is already so proficient at copying a particular artist’s work it won’t be long before filmmakers need to protect themselves from plagiarism.

article here 

There could even be a need right now to copyright camera moves, editing choices, color palettes, lighting schemes, or compositions because there is nothing to prevent an AI from entirely generating a new gangster movie in the style of Martin Scorsese or a sci-fi film that looks and feels like it has come from Stanley Kubrick.

On the other hand, there will be some in Hollywood no doubt calculating that if an AI could perfect a hit movie without having to pay for the fuss, the time, all the micro-decision making and risk that human talent brings it’s a price worth paying.

This is not idle speculation; the debate about artistic infringement by algorithm has become a hot one in the art world.

Swedish artist Simon Stålenhag is among those sounding a warning. “AI basically takes lifetimes of work by artists, without consent, and uses that data as the core ingredient in a new type of pastry that it can sell at a profit with the sole aim of enriching a bunch of yacht owners,” he tells Wired’s Will Knight.

Stålenhag’s style was recently used to create images on the text-to-image AI Midjourney by academic Andres Guadamuz in an apparent attempt to draw attention to the legal issues surrounding AI-generated art.

Stålenhag was not amused. In a series of posts on Twitter, he said that while borrowing from other artists is a “cornerstone of a living, artistic culture,” he dislikes AI art because “it reveals that that kind of derivative, generated goo is what our new tech lords are hoping to feed us in their vision of the future.”

The dawn of a new era of AI art began in January 2021, when OpenAI announced DALL-E, a program that used recent improvements in machine learning to generate simple images from a string of text.

In April this year, the company announced DALL-E 2, which can generate photos, illustrations, and paintings that look like they were produced by human artists. This July, OpenAI announced that DALL-E would be made available to anyone to use and said that images could be used for commercial purposes.

 “As access to AI art generators begins to widen, more artists are raising questions about their capability to mimic the work of human creators,” Knight says.

Digital artist David OReilly, for instance, tells Knight that the idea of using AI tools that feed on past work to create new works that make money feels wrong. “They don’t own any of the material they reconstitute,” he says. “It would be like Google Images charging money.”

But it’s not clear if the legal framework is strong enough to protect an artist’s work from AI-generated imitation.

In a blog post, Guadamuz argued that lawsuits claiming infringement are unlikely to succeed, because while a piece of art may be protected by copyright, an artistic style cannot.

 

Lawyer Bradford Newman tells Knight, “I could see litigation arising from the artist who says ‘I didn’t give you permission to train your algorithm on my art.’ It is a completely open question as to who would win such a case.”

In a statement to Wired, OpenAI defended DALL-E 2, saying that the company had sought feedback from artists during the tool’s development.

“Copyright law has adapted to new technology in the past and will need to do the same with AI-generated content,” the statement said. “We continue to seek artists’ perspectives and look forward to working with them and policymakers to help protect the rights of creators.”

Painted art, like motion pictures or literature, evolves and builds upon everything and everyone that has gone before it. Imitation could be homage or pastiche. Brian de Palma’s work, including Body Double and Dressed to Kill, was heavily influenced by Alfred Hitchcock. Hitchcock’s film Psycho has been recreated shot-for-shot by Gus Van Sant, in color. The infamous shower scene in Psycho is credited to Hitch, but may have been designed by Saul Bass. There is no clear line between imitation as flattery and straight out plagiarism.

As AI gets more and more sophisticated to produce longer-form narrative video, including deep fake or CG actors, the dividing lines will increasingly blur.

Short film The Crow shows just how far text-to-video has come.

 One worry for Hollywood is that while AIs might make certain types of production cheaper to churn out, the same technology could easily be in the hands of anyone. DALL-E 2 and Midjourney, for example, are simple enough to operate by just typing (or saying) a series of simple words.

On the other hand, there is an argument that AI art should not be given the same equivalence as art generated by a human. It should matter, the argument goes, that a piece of content pertaining to mean something has been created by people who have actually lived the experience.

If AI video is inevitable perhaps the AI owner, the AI prompters, and all of the AI’s data-set training wheels should be credited in the titles?

No comments:

Post a Comment