In 2024, government regulators worldwide are poised to address critical questions posed by the rapid advancement of generative artificial intelligence (AI) technology. These questions, centered on issues such as licensing for training models on copyrighted material and the copyright protection of AI-generated outputs, are pivotal in shaping the future of AI adoption and the intellectual property landscape.
Licensing Dilemmas: A Global Perspective
Will Licensing Be Required?
The looming question for tech giants like OpenAI Inc. and Stability AI Ltd. revolves around the necessity of licenses for training AI models on copyrighted material. The outcome of this query could significantly influence the investment priorities of the technology industry globally and impact how creative industries safeguard their intellectual property.
Copyright Protection for AI Outputs
Equally critical is the inquiry into whether the outputs of AI models will receive their copyright protections. The answers to these inquiries hold immense implications for industries like film and graphic design, dictating where they establish their new offices to safeguard their intellectual property.
Want to know if you’re earning what you deserve? Find out with LawCrossing’s salary surveys.
The Urgency for Formal Regulations
The AI industry awaits formal regulations to establish clear guardrails, and Microsoft Corp. has explicitly tied regulatory clarity to progress. According to Burton Davis, Microsoft’s Deputy General Counsel for IP, countries providing the most fantastic clarity will witness the greatest adoption of responsible AI technology. Microsoft emphasizes that, without regulatory clarity, the next phase of AI development is in jeopardy.
Global Regulatory Perspectives
United States: Stricter Stance on AI Copyright
In the United States, the stance on AI copyright is strict. Courts have ruled that only human authors can receive copyright protection. The US Copyright Office demands disclosure of AI used in registrations, rejecting applications for AI-generated images, contributing to concerns about stifling creativity.
China: Embracing AI Copyright
Contrastingly, the Beijing Internet Court ruled that an AI image created with Stable Diffusion could receive copyright protection. This approach differs significantly from the US policy, potentially disadvantaging the US creative industry.
United Kingdom: Legacy Law Supports AI Copyrightability
The UK’s Copyright, Designs, and Patents Act of 1988, crafted to protect computer-generated works, could inadvertently support the copyrightability of modern AI-generated works.
Copyrighting AI Content: A Global Conundrum
Legal Battles Over AI Art
The legal battles over AI-generated art are intensifying. While some argue that AI art generators are tools akin to cameras, others view them as autonomous programs making artistic decisions without human involvement.
Divergent International Policies
International policies on AI copyright vary widely. With its stricter stance, the US contrasts with more lenient approaches in China and potential support in the UK. Attorney Ryan Abbott emphasizes concerns about rejecting works with human creativity due to a perceived association with artificial intelligence.
Training AI: Global Perspectives on Legality
Fair Use Debate
The legality of training AI models on copyrighted content remains a global debate. In the US, courts are yet to determine if this falls under the “fair use” standard, while the UK’s High Court of Justice allowed a copyright case against Stability AI to proceed.
Global Variances in Exemptions
Countries like Japan, Singapore, and Israel have proactively created exemptions for AI training. Japan explicitly allows data scraping for training machine learning models, while Singapore’s 2021 copyright law includes exemptions for computational data analysis.
European Union’s Stance
The draft version of the European Union’s forthcoming AI Act includes exemptions for text and data-mining operations but allows rights holders to opt-out.
Microsoft’s Optimism
Microsoft’s Chief Legal Officer, Shelley McKinley, expresses optimism that jurisdictions will recognize the principle that AI training extracts “uncopyrightable elements” from data, protecting the ability to train on public data.
Don’t be a silent ninja! Let us know your thoughts in the comment section below.