Artificial intelligence is reshaping markets at breakneck speed – but it may also be creating new ways for competitors to collude, according to Norton Rose Fulbright’s antitrust specialists.
In a recent article, Marianne Wagener, the head of the firm’s antitrust and competition team, and candidate attorney Adriaan Lourens, say regulators face a growing dilemma: how to encourage innovation while curbing anti-competitive behaviour in increasingly complex digital markets.
Wagener and Lourens explain that AI tools, particularly when trained on similar data or used by multiple competitors, can end up producing aligned pricing or market strategies without any explicit agreement or intention from the firms themselves. Over time, what looks like independent decision-making could drift into what competition law calls a “concerted practice” – which is prohibited.
AI could cause collusion even if companies don’t mean for it to happen.
The legal backdrop
They pointed out that the Competition Act prohibits price fixing, market allocation, and bid rigging – and that no formal agreement is needed to establish collusion.
“Importantly, the Act does not require a formal or express agreement to establish collusion,” they wrote. Even concerted practices – co-ordinated conduct that replaces independent action – are outlawed.
At first glance, the use of AI tools by individual firms may appear harmless. But the authors cautioned that AI could still generate co-ordinated strategies, particularly if systems are trained on the same data or built on similar models. The risk rises further if rivals use the same AI tools, because this can lead to pricing being indirectly aligned through a shared system.
“Conduct that initially appears independent and lawful may, over time, develop into a concerted practice,” they warned.
Information-sharing risks
The experts also flagged the danger of AI inadvertently enabling firms to share sensitive information. Since AI models rely on massive datasets, it is often unclear what data is included – and whether it contains confidential pricing or customer information from competitors.
“If this includes sensitive information from competing firms … the risk of unintended collusion becomes significantly higher,” they cautioned.
Gaps in regulation
South Africa currently has no dedicated AI competition law framework, and the Competition Commission has not issued specific guidance. By contrast, the European Union has rolled out its AI Act, which expands oversight powers, although it does not directly deal with AI-driven collusion.
In the United States, regulators are probing whether common pricing algorithms could amount to unlawful concerted action. These investigations are ongoing, but no binding legal precedents have yet been set.
What companies should do
Given the uncertainty, Wagener and Lourens urge businesses to put safeguards in place. Firms, they said, should thoroughly vet AI vendors, include compliance clauses in service agreements, and conduct regular audits of their tools.
They also emphasised the importance of controlling how data is shared. “Public availability does not necessarily render information non-sensitive. Firms must carefully assess both the nature and extent of any disclosure, limiting it to what is strictly necessary.”
The pair concluded with a clear warning: “AI offers significant advantages but also introduces complex legal risks. Firms must carefully manage the AI tools they use and put strong compliance systems in place to reduce competition law risks.”
