HARDIN RATSHISUSU | Regulating AI before it regulates us

Artificial intelligence (AI) is reshaping markets, economies and societies faster than regulators have ever had to respond in previous technological revolutions.

AI is promising great productivity, innovation and consumer choice, but it also has big risks — concentration of market space and consumer abuse, and extreme inequality.

The question is no longer whether AI should be regulated but whether regulation can keep pace with the technology itself.

Consumer protection and competition authorities all over the world are dealing with this new reality. At the recent UN Conference on Trade and Development (Unctad) meetings of the Intergovernmental Group of Experts on consumer protection and competition law and policy, the topic of product safety, food value chains, digital markets, and AI and how it will affect competition and consumer welfare were among the topics of discussion.

The timing was significant. In the same week, world leaders, policymakers, scientists and business leaders gathered for the world’s first UN Global Dialogue on AI Governance. The convergence of these discussions reflects an increasing global consensus — AI is transforming markets and society at an unprecedented pace. The challenge is to embrace that opportunity and ensure it doesn’t reproduce and exacerbate inequality.

For consumer protection and competition regulators, the urgency is clear. New markets are emerging, existing business models are changing and regulators themselves are looking at AI-driven investigative tools to better analyse and monitor markets.

However, the more difficult task is reacting quickly enough to the new risks AI is creating. Consumers are now far more susceptible to deepfakes, misinformation, disinformation, scams, fraud and misleading online advertising, all compounded by the misuse of AI.

The scope of harm is immense. Even the most educated consumers can’t discern genuine content from sophisticated manipulation. At the same time, AI helps bad actors target the most vulnerable consumers with unprecedented precision and make exploitation easier and more widespread.

Competition is also of great concern. Digital markets such as search engines, social media and e-commerce show how quickly markets can become concentrated in the hands of a few firms. First-mover advantages, network effects and control over major infrastructure often create barriers to entry that limit competition and consumer choice.

We do not need to revisit the many regulatory interventions that have emerged from digital markets. It is enough to realise their evolution has not always brought competitive or socially optimal outcomes.

AI appears to be following a similar trajectory. The market is converging around a relatively small number of large language model (LLM) developers with advanced deep-learning capabilities.

Top platforms include ChatGPT by OpenAI, Claude by Anthropic, Gemini by Google, Copilot by Microsoft, Meta AI by Meta and DeepSeek by Zhejiang.