[GENERATIVE IA - IN SEARCH OF THE ORIGIN AND THE ORIGINALITY]

[WORK - ORIGINAL - CREATION - IA - INTELLECTUAL PROPERTY - COPYRIGHT]

🚨With the advent of artificial intelligence systems (AIS), it has become necessary to define and regulate the relationship between the human author-creator, generative AI and the designer of the latter.

⚖️ On 12 September 2023, a bill (no. 1630) was tabled by members of the French National Assembly with a view to promoting innovation and artistic diversity.

Its main aim is to regulate the conditions for exploiting works generated by AIS, and to ensure fair and equitable remuneration for their use, for the benefit of the authors and artists affected.

The bill n°1630 suggests to :

- Amend article L.131-3 of the French IP Code: authorisation of authors would be required prior to the integration of their works by an AIS;

- Complete article L321-2 of the French IP Code: in the absence of direct human intervention in the creation of the work in question, the ownership of this work would revert to the creator/designer of the AI;

- Complete of article L121-2 of the French IP Code: addition of the mandatory mention ‘work generated by AI’ to respect for the author's moral rights; and

- Complete article L121-2 of the French IP Code: establishment of a tax for the benefit of the entity responsible for collective management, if the origin of pre-existing works exploited by AIS is not determinable.

⚠ On December 27, 2023, New York Times filed a lawsuit in Manhattan federal court against OpenAI and Microsoft. The case concerns OpenAI and Microsoft's unauthorised exploitation of millions of New York Times articles to train ChatGPT and Copilot AI technologies. The latter were able to generate texts that ‘recite the content of The Times word for word, summarising it as closely as possible and imitating its style’.

The damage claimed by New York Times, estimated at ‘billions of dollars’, is the alteration of the relationship with the readership and the deprivation of the publisher of revenues from subscriptions, advertising, affiliation and paid licences to exploit its content. The New York Times is seeking the withdrawal of content exploited in breach of copyright.

OpenAI argues that it would be ‘impossible to create tools such as its chatbot without accessing copyrighted material’ and that ‘legally, copyright does not prohibit training’. It invoked the US doctrine of ‘fair use’ of copyrighted material without authorisation in certain circumstances.

On March 4, 2024, Microsoft filed its case against Open AI in the U.S. District Court for the Southern District of New York. Microsoft argued that large language models (LLMs) had not supplanted the relevant market.

(The New York Times Company v. Microsoft Corporation, 1:23-cv-11195, (S.D.N.Y. Mar 04, 2024) ECF No. 65)

⚠ On 8 March 2024, Nvidia was sued by novelists Brian Keene, Abdi Nazemian and Stewart O'Nan in an American class action suit filed in federal court in San Francisco. The authors allege that in order to train NeMo (Nvidia's AI platform) to simulate ordinary written language, a database of approximately 196,640 books (including their works) was exploited.

All of these proceedings are still pending before the respective courts.

#IA #AI #original #creation #work #copyright #PI #IP #property