In a bid to ensure competition and consumer protection, the UK’s Competition and Markets Authority (CMA) has launched an initial review of AI foundational models, such as large language models (LLMs) and generative AI models that power OpenAI’s ChatGPT, Microsoft’s Bing Chat, OpenAI’s DALL-E and Midjourney. ChatGPT is an AI-powered chatbot that is trained to answer questions.
The review follows the UK government’s white paper in March, which asked regulators to consider how the innovative development and deployment of AI can be supported with governance principles such as safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; and contestability and redress. Accordingly, the CMA’s review seeks to understand how foundation models are developing and assess the best guiding conditions and principles for future model development and usage.
Stanford University’s Human-Centered Artificial Intelligence Center’s Center for Research on Foundation Models coined the term “foundation models”. Foundation models are AI systems that focus on training one model on a wide range of data and adapting it to many applications. This opens up the possibility of generating human-like responses to natural language inputs, with the potential to transform much of what people and businesses do, including in the music industry. Examples of generative AI models include OpenAI’s Jukebox, which can generate lyrics, compose melodies and create entire songs, as well as ByteDance’s Jukedeck, and Shutterstock’s Amper Music.
However, the use of foundation music in music also raises concerns about copyright and ownership, as seen in the recent “fake Drake” and AI-generated music controversy. Investors are also wary of the potential risks associated with AI, as demonstrated by venture capital firm Sound Ventures’ recently closed Sound Ventures AI Fund at nearly $240 million.
The CMA is seeking views and evidence from stakeholders until June 2. Its review findings will be published in September.
