Algorithm of the night
Creativity and copyright in the face of artificial intelligence
Fourth-year Law student Max Bicknell spent last summer thinking about what the future holds for human creativity and copyright law as part of his summer research scholarship project.
For three months Max worked alongside a team of policy advisers at the Ministry of Business, Employment and Innovation (MBIE) investigating the potential legal and commercial implications of artificial intelligence (AI) and machine learning for New Zealand businesses and artists. Professor Graeme Austin from Victoria’s Faculty of Law provided academic supervision.
Max was attracted to the project because of his fascination with the dynamic world of technology and AI. “I’ve definitely been keeping up with the current state of AI—I find it interesting because it’s moving so fast.”
Max’s research scholarship culminated in a report which he presented to the team at MBIE. The report provided background information about current uncertainties in AI copyright issues, what other countries are doing in this area, and what the possible options are for New Zealand.
“There are two major legal issues when it comes to copyright and AI, which can be broadly described as the ‘input issue’ and the ‘output issue’,” says Max.
He says the output issue is simply the question of when AI creates a work, such as an image or a song, who owns it—the programmer, the user of the AI, or perhaps even the AI itself?
While this is an important consideration, Max discovered that the ‘input issue’ of how an algorithm learns is even thornier in terms of copyright law.
A central part of developing algorithms is that they need to copy data in order to learn, says Max. “If my algorithm comes across your copyrighted photo of a cat on the internet when it’s learning how to recognise cats, it will scan the photo and get data from it, which raises the question: is that copyright infringement of your work?
“My research found that no matter which way you spin it, it’s an infringement because my machine would have to copy that photo—however briefly.”
Max’s report concluded that New Zealand’s current legislative set-up for copyright is not fit-for-purpose, and that the challenge for policy makers is to make legislation broad enough that it will capture future changes. “Technology moves a lot faster than the law,” says Max.
He says a further challenge lies in finding a balance between protecting intellectual property and promoting the development of new technologies. “The stricter the copyright laws, the better you’re protecting producers—but you risk hampering development.”
Despite these obstacles, Max thinks New Zealand is doing relatively well in the area of copyright law compared with other countries. “We’ve got a provision for ‘computer-generated works’ in our Copyright Act, so we’re a bit further ahead than some other countries.”
However, since this law was passed, there have been dramatic advancements in machine intelligence and the capacity of AI to create new works—“there are already poems and symphonies that have been created by AI, and it won’t be long before AI can do most of what humans do creatively.”
The question of who gets paid royalties from machine-generated works is a timely one, as it was recently reported that music streaming giant Spotify has hired François Pachet, the former Director of the Sony Computer Science Laboratory in Paris, who oversaw the creation of the first known AI-composed pop songs.
Max is a singer and musician himself, and has been the musical director of the annual Law revue, but he says he doesn’t feel too threatened by the rise of machine creativity—“I think there will always be a desire for human expression.”
As well as providing ample food for thought, Max says the summer scholarship opened his eyes to new career paths. “I’d never really considered working in the public sector before, but the team I worked with at MBIE were doing some really cutting-edge stuff around copyright, technology and the creative sector.”
MBIE recently announced that it will be doing a full-scale review of the Copyright Act, and Max’s report will contribute towards this work as policy makers explore options for a revised Act.
Max’s supervisor Professor Austin says, “It was great to see how much Max’s research and writing skills developed during his summer scholarship experience—and it was terrific for him to have the opportunity to work with MBIE on such important issues.”
Having spent longer than most 22-year-olds contemplating the future of human endeavour and creativity, what does Max think is in store for us?
“There might be a point where we need to regulate AI—whether it’s to protect jobs from automation or to protect human creativity in terms of copyright—but it will have to be an international effort,” says Max.
“No one can really predict what will be the next big thing in ten years. On some days I’m terrified and other days I think ‘this is going to be awesome’.”