Photo Credit: Gonzalo Poblete
OpenAI, Microsoft, and Google: Negotiating Data Deals with News Outlets for AI Development
According to reports from reputable sources like the Financial Times, OpenAI, Microsoft, and Google have been engaging in discussions with news executives. The purpose of these deliberations revolves around addressing potential copyright challenges that arise from the use of AI-generated content. Notably, large language models such as ChatGPT, Midjourney, and Stable Diffusion are trained on existing content. An ongoing legal dispute between Getty Images and Stable Diffusion further underscores the presence of copyright-related complexities.
Subscription-Style Fee: A Potential Solution to Copyright Issues
The Financial Times reveals that influential news organizations including News Corp, The New York Times, The Guardian, and Axel Springer have been exploring the idea of adopting a subscription-style fee for providing access to their content to develop AI models. The outcome of such a deal could serve as a guiding framework for handling similar copyright concerns related to generative AI. While the exact pricing structure is yet to be determined, early indications suggest that it may range from $5 million to $20 million per year for acquiring access to the content.
Adapting the Music Industry Model for AI Generated Content
Discussions between these companies and news outlets also explore the possibility of creating a “quantitative model” that operates similarly to the system employed by the music industry. In the music industry, record labels receive compensation whenever their tracks are played on radio stations, night clubs, or streaming services. News organizations aspire to establish a similar model where content owners are remunerated each time their information is accessed. To implement such a model, companies would need to disclose their current usage of media content, a measure which these organizations are currently unwilling to take.
OpenAI’s Potential Response to EU Regulations
The CEO of OpenAI, Sam Altman, has gone so far as to suggest that his company might halt operations in the European Union (EU) if the region enforces legislative regulations governing its activities. The proposed EU AI Act would compel entities like OpenAI to disclose copyrighted material employed in training their AI systems. Dutch MEP Kim van Sparrentak, who actively participated in the drafting of the current EU AI Act, commented that if OpenAI fails to comply with fundamental data governance, transparency, safety, and security requirements, its systems would not be deemed suitable for the European market.
Editor Notes: Analysis and Opinion
It is evident from recent developments that data deals between AI companies and news outlets are becoming prevalent. While comprehensive agreements on copyright and compensation are yet to be reached, the negotiations signify a concerted effort by industry players to address these challenges collectively.
On one hand, news organizations are rightfully seeking fair compensation for their content’s usage, as AI-generated content draws heavily from their intellectual property. Implementing a payment model similar to what is observed in the music industry seems like a sensible approach to ensure proper remuneration.
On the other hand, AI companies like OpenAI, Microsoft, and Google also face the task of navigating legal complexities while advancing their AI technologies. Balancing data governance, transparency, safety, and security requirements is crucial for these companies to maintain trust with the public and regulatory bodies.
The resolution of copyright matters within the field of generative AI is likely to set future precedents and shape the landscape of AI development. As the industry progresses, finding mutually beneficial solutions that benefit both news organizations and AI companies will be pivotal.
Discover More AI News on GPT News Room
For the latest news and insights in the field of AI and technology, visit GPT News Room. Stay informed and stay ahead with our in-depth coverage and expert analysis.
from GPT News Room https://ift.tt/yTwJe94
No comments:
Post a Comment