UK Antitrust Watchdog Announces Initial Review of Generative AI

Image credit: Leon Neal/Getty Images

Well it was fast. The UK competition watchdog has released an early review of “AI foundational models” such as OpenAI’s ChatGPT and Microsoft’s Large Language Models (LLM) that underpin Microsoft’s New Bing. Generative AI models that power AI art platforms such as OpenAI’s DALL-E and Midjourney may also be of interest.

The Competition and Markets Authority (CMA) said it would consider competition and consumer protection considerations in the development and use of foundational models for AI. It will best guide the development of the underlying models and their use in the future. ”

The review is proposed to be published in “early September”, with a June 2nd deadline for interested stakeholders to submit responses to inform them of the work.

“The foundational models that have emerged over the past five years, including large-scale language models and generative artificial intelligence (AI), have the potential to transform much of the behavior of people and businesses. To ensure that AI innovation continues in a way that benefits the economy, governments should: [CMA], how innovative development and deployment of AI can be supported against five overarching principles: safety, security, and robustness. Appropriate transparency and explainability. fairness; accountability and governance; CMA said in a press release:

The Foundation Model Research Center at Stanford University’s Center for Human-Centric Artificial Intelligence is known for coining the term “Foundation Model” in 2021. many applications.

“The development of AI involves many important issues, not just how the market works, but also safety, security, copyright, privacy, and human rights. As such, this first review will focus on the issues that the CMA is best suited to address — what impact the development of AI-based models for competition and consumer protection could have. is there?” CMA added.

In a statement, its CEO Sarah Cardell also said:

AI has been on the radar for some time, even though it has exploded into the public eye over the past few months. It is a rapidly evolving technology that has the potential to transform the way businesses compete and drive substantial economic growth.

It is important to ensure that UK businesses and consumers have easy access to the potential benefits of this innovative technology, while protecting people from issues such as false or misleading information. Our goal is to help develop this rapidly expanding new technology in a way that ensures an open and competitive marketplace and effective consumer protection.

Specifically, the UK competition regulator says its first review of AI’s underlying models will:

• Examine how the underlying model and the competitive market for its use evolve
• Explore what opportunities and risks these scenarios pose to competition and consumer protection
• Create guidelines to support competition and protect consumers as AI-based models evolve

It may be too early for antitrust regulators to review such fast-moving emerging technologies, but the CMA is acting on government directives.

An AI white paper released in March indicated ministers’ intentions to avoid setting bespoke rules (or oversight bodies) to govern the use of artificial intelligence at this stage.but Ministers said they expected existing UK regulators (including the directly named CMA) to issue guidance to promote the safe, fair and accountable use of AI. rice field.

The CMA says the initial review of the basic AI model is in line with the white paper’s instructions. The government said existing regulators are in a position to conduct a “detailed risk analysis” to carry out potential enforcement. Unfair and unexplained applications of AI’s existing power.

Regulators also cite their core mission of supporting open and competitive markets as another reason to consider generative AI now.

Notably, the competition watchdog is expected to gain additional powers to regulate big tech in the coming years under a plan that was shelved by Prime Minister Rishi Sunak’s government last month. Preliminary reform aimed at the market power of the digital giants.

The CMA’s digital markets unit, which has been operating in shadow since 2021, is expected to (eventually) gain legislative power and apply aggressive ‘pro-competitive’ rules in the coming years. . Has a “Strategic Market Status” (SMS). So we can speculate that the provider of a strong underlying AI model may eventually be determined to have SMS. That means you can expect to face bespoke rules on how you must operate against competitors and consumers in the UK market.

The ICO, the UK data protection watchdog, is also eyeing generative AI. This is another existing watchdog tasked with giving governments special attention to AI under a program of context-specific guidance to guide technology development through the application of existing laws.

In a blog post last month, the ICO’s Executive Director of Regulatory Risk Stephen Almond offered some tips and caveats to generative AI developers regarding compliance with the UK’s data protection regulations. “Organizations developing or using generative AI should consider their data protection obligations from the start and adopt a data protection by design and default approach,” he suggested. “This is not an option. If you are processing personal data, it is the law.”

Meanwhile, in the English Channel of the European Union, parliamentarians are deciding on a set of rules that will likely apply to generative AI.

Negotiations are underway towards the final text of the EU’s upcoming AI rulebook, but now, through amendments to the risk-based framework for regulating the use of AI, which the bloc published in draft over two years ago, The focus is on how to regulate the base model.

It remains to be seen whether the EU co-legislators will end up with what is sometimes called general-purpose AI. However, as we recently reported, parliamentarians are promoting a layered approach to addressing safety issues using a basic model. The complexity of responsibilities across the AI ​​supply chain. It can also address specific content issues related to generative AI, such as copyright.

In addition, EU data protection laws already apply to AI. Also, privacy-focused research into models like ChatGPT is ongoing within the block. Including Italy, where interventions by local watchdogs last month led OpenAI to scramble a series of privacy disclosures and controls.

The European Data Protection Board also recently established a task force to support coordination among various data protection authorities on investigations into AI chatbots. Other members investigating ChatGPT include the Spanish privacy watchdog.