Publishers seeking to apply existing copyright laws to ChatGPT and other AI-generated works may face an uphill battle.
Copyright law, such as the European Union’s Copyright Directive, guarantees media outlets the right to seek fair compensation from social media sites and search engines that use their work. But the press is struggling to apply these laws to generative AI, and regulators are taking notice.
A confusing new reality
Copyright law has been around since the 18th century, but generative AI has found a tough new test. Both outlets and policy makers now have to grapple with the challenge of chatbots using data without authorization or compensation.
Petra Wikström, director of public policy at Sibsted, Scandinavia’s largest media group, is among those concerned about the technology and its implications.
“We don’t enter the AI discussion in fear, but of course there are issues to consider, and copyright is one of them,” said Wikström. Politico just a while ago this month.
in the case of Generative AI, the problem is to identify whose copyright may have been infringed. Facts themselves are not copyrightable, and data mining is generally an exception to copyright law. But if an article written by a human is completely copied into an AI database, it certainly could be in violation of copyright law.
The question is, what exactly is collected and stored in these AI databases?
Iacob Gammeltoft, Policy Manager at News Media Europe (representing 2,400 people) european news outlet) admits that AI has presented the organization and its members with an almost unsolvable copyright conundrum.
“Our work has copyright protection, but the question is how can it be exploited? Enforcement is the main issue,” Gameltov said.
One of the major obstacles to the discussion of generative AI is copyright.
This is especially true for the EU. EU privacy laws have set the direction of programming for many years.
— Ant 💀 Currently writing a novel (@AGramuglia) April 18, 2023
keep away from the grass
Data mining may be one of the major exceptions to copyright law, but that doesn’t mean publishers have to acquiesce to it.
Copyright holders may place machine-readable code on their website to warn against data miners. A digital version of the “lawn fort” sign. Some jurisdictions are currently considering whether the application of these no-go signs to chatbot technology might provide some sort of solution.
The French publisher of the GESTE association is considering whether digital invasiveness laws can be used in this way and whether it is the right direction.
There are certainly voices who agree with the need to take action. The problem, however, is determining whether these no-go signs were violated by AI.
As Gammeltoft admits, “There’s no way to be 100% sure until you check the training database. [has] Used for ChatGPT training. ”
Without the full transparency of the ChatGPT database, publishers are playing the guessing game or doing expensive and time-consuming research work.
Brief description of AI copyright
Copyright laws vary by jurisdiction. In common law countries such as the UK, the focus of copyright law is to protect creators financially. Continental Europe emphasizes the natural rights of creators. In either case, copyright law exists to protect the author of a work.
Copyright is assumed without further action when an author creates a work. If a third party uses a substantial portion of a copyrighted work without the author’s permission, it may violate the copyright of the original author and entitle them to some form of monetary compensation. I have.
When copyright differs from person to person AI It is in the human head and in the AI database. If a human remembers part of a copyrighted work, no infringement occurs. But if the AI stores a significant portion of the copyrighted work in its database, infringement may have occurred.
Unfortunately the problem is proving it.without access to AI‘s database makes it very difficult to identify cases of copyright infringement, and that is where the difficulty lies. Whatever action the EU takes next, the road ahead looks very daunting when it comes to cracking down on US-based tech companies.