Jerry Crisci’s Post

The UK Intellectual Property Office (UKIPO) led discussions to create a voluntary code of practice on copyright and Generative AI ("Gen AI") but failed to reach an agreement. The working group could not reach consensus on the use of copyright-protected works to train Gen AI models. This outcome disappointed both the creative and technology industries, as they sought clarity on protecting works and developing future technologies. The creative industry expressed concerns about the large-scale use of works to train Gen AI models, while AI developers emphasized the need for access to diverse datasets. The UK Government plans to lead further engagement with right holders and AI sectors to find an effective approach for partnership. This announcement coincides with a consultation series initiated by the UK Information Commissioner's Office on how UK Data Protection Laws should apply to the development and use of Gen AI.

 The significance of transparency in the development and use of Generative Artificial Intelligence (AI), particularly in relation to copyright issues are many, below is a summary of the attached article, that you can read for yourself. Here are the key points: Transparency in Training Data: There is an emphasist in regards to the need for AI companies to be open about the materials used to train their AI systems. Output Labeling: The article suggests a necessity to inform the public when content is generated by AI, though this topic is reserved for a future blog post. Copyright Infringement Concerns: The article highlights the increasing number of copyright infringement lawsuits against AI companies and the secrecy surrounding the training data. Importance of Transparency: It argues that transparency can ease the legal process for copyright owners and support ethical AI development. The article also gives several examples of lawsuits where copyright owners have shown that the unauthorized use of their works in training AI systems. It concludes by advocating for transparency rules to protect copyright owners and ensure responsible AI development.

“One of the frontiers of innovation in AI should be in techniques and business models to enable the kind of flourishing ecosystem of content creation that has characterized the web and the online distribution of music and video. AI companies that figure this out will create a virtuous flywheel that rewards content creation rather than turning the industry into an extractive dead end.” Some good thoughts in the below O’Reilly piece that lays out the reasoning behind their business model innovation. But internet sourcing is a naïve premise as it flips the principle of informed consent and fatally ignores massive power asymmetries of bad actors and amateurs: In the real world, any popular content by definition ends up online via third parties without the rightsholder’s permission, regardless if they are even online, aware or in a position to police it. It is put there by actors with no incentive to apply any opt-out mechanisms, and as the Internet is ultimately anonymous, no disincentives. “Information wants to be free.” And any popular creative activity generates enough non-opted amateur content that models built only on it will hit replacement quality, effectively killing any incentive to learn the skill; market incentives first, effort incentives second. Opt-out registries are necessary but not sufficient. Fundamental legal principles must be upheld. #degenerativeai not built on consent inevitably kills it substrate.

Nicholas Thompson When Google provides "answers" that are an assemblage of dozens of articles and work product, should Google pay all of the original authors? 2. When someone reads a book that is given to them by someone else, should the author be paid twice? 3. Just because some copyright is "old", does that mean the original authors or their heirs should not receive compensation? 4. Your question turns on "derivative". There is nothing original under the sun. 5. What if the work is found to be 49% "derivative" and 51% new, should the original author get 49% of the payment?

CCC understands how important it is to enable the development of ethical, reliable, and trustworthy AI systems to further business and research goals. At the same time, there are meaningful risks associated with some AI applications. AI systems trained on poor quality materials can generate bad science, promote misinformation, and lead to harmful results. AI-related advancements will be strengthened by relying on responsibly sourced copyrighted works. Catherine Zaller Rowland and Babis Marmanis will give a great talk on March 20 to Pistoia Alliance colleagues on this important topic.   

AI act: expressly avoiding potential IP forum shopping Recital 106 of the AI Act: Any provider placing a general-purpose AI model on the Union market should comply with this obligation [to comply with Union law on copyright and related rights], REGARDLESS of the jurisdiction in which the copyright-relevant acts underpinning the training of those general-purpose AI models take place. This is necessary to ensure a level playing field among providers of general-purpose AI models where no provider should be able to gain a competitive advantage in the Union market by applying lower copyright standards than those provided in the Union.

Our latest podcast episode with Christopher Young and Briar Richardson on intellectual property in the context of generative AI. In this episode, Partner Chris Young is joined by Senior Associate Briar Richardson to discuss intellectual property in the context of the rapidly evolving landscape of generative artificial intelligence (AI). They discuss the absence of AI specific IP legislation in New Zealand and outline existing copyright and patent laws and how these laws may impact use of generative AI. Touching on current legal cases including Getty Images in the UK and GitHub Copilot in the US, Chris and Briar outline the jurisdictional challenges and some of the key legal issues before the Courts. The conversation also extends to some business risks associated with generative AI and provides practical precautions for organisations. #ai #aiact #intellectualproperty #copyright #patents

If you would ask me in 10 years from now "What was THE hot topic among IP and IT lawyers in 2024?" the answer will be simple: generative AI or GenAI! Indeed, the use of generative AI raises many interesting questions, especially in the field of copyright. In this second blogpost of our 'AI & IP' series, together with my colleagues Alexis Fierens and Emma Stockman we are happy to demystify the first and most basic question: are AI-generated works eligible for copyright protection? Why would you ask ChatGPT if you can also read our contribution?

Exciting times in the AI world! A shift in accountability for copyrighted materials is on the horizon. Congressman Adam Schiff has introduced a new bill, the Generative AI Copyright Disclosure Act, which could have huge implications for the AI sector. It requires AI companies to reveal copyrighted materials used to build their generative models at least 30 days ahead of launching the AI tool, or face sanctions. The pressure is on to reveal trade secrets - making industry giants thinking twice about how they use creative property. No outright banning of copyrighted material is on the cards, but the bill sets a significant burden on companies to disclose the array of works they use, data traditionally kept under wraps. This bill could create transparency in the industry, prevent misuse of copyrighted content, and ensure credit is attributed where it's due. How will this legislation affect the future of AI and new tool creation, and does this herald new levels of transparency industry-wide? #AI #ArtificialIntelligence #CopyrightLaw #OpenAI #AITransparency