AI chatbots and competition law

The rise of conversational AI agents, from simple chatbots to sophisticated assistants capable of autonomous commerce, is transforming how businesses and consumers interact with digital services.  

As this technology is evolving, competition authorities globally are scrutinising whether the largest tech platforms may be able to exercise control over access to emerging AI markets.  

The European Commission’s (EC) ongoing antitrust investigation into Meta’s WhatsApp policy, and the swift regulatory interventions that have followed in Italy and Brazil, provide an insight into the competition risks associated with conversational AI and underscore a broader regulatory theme: global regulators are intensifying scrutiny at the intersection of emerging AI markets and platform market power. We are also seeing the use of interim measures as an enforcement tactic in this space – a sign that regulators want to move quickly to unlock barriers to competition in AI markets.

In this article we take a closer look at global antitrust enforcement in relation to WhatsApp’s new AI policy.

What is WhatsApp’s new AI policy for business users?

In October 2025 Meta updated its WhatsApp Business Terms to regulate how businesses can deploy AI-based chatbot services on WhatsApp. The policy’s key feature is a restriction on third-party AI chatbots on WhatsApp and came into force from 15 January 2026 for AI providers already present on WhatsApp, whilst for new AI providers the policy took effect on 15 October 2025.

The new terms ban businesses from using the WhatsApp Business API if the primary service they offer to users is an AI chatbot or assistant.  Businesses can continue to use AI-driven features on WhatsApp as long as the AI is auxiliary to their main service. This prompted concerns that it would exclude rival chatbot providers from the platform as the only AI assistant that can be used on WhatsApp’s Business Platform is its own “Meta AI” tool.

The timing is significant as Meta AI had just been rolled out to WhatsApp users in late 2025.

Coordinated Global Enforcement Response

European Commission Investigation (4 December 2025)

On 4 December 2025 the EC launched an investigation probing whether Meta’s new WhatsApp policy might constitute an abuse of its dominant market position in breach of Article 102 of the TFEU.

On 9 February 2026 the EC issued a Statement of Objections to Meta informing them that “this policy change appears at first sight to be in breach of EU competition rules”. The Commission has confirmed that it intends to use interim measures to prevent Meta’s policy change from causing irreparable harm in the rapidly growing market for AI assistants.

The EC’s preliminary conclusions are that:

• Abuse of Dominant Position: Meta, through WhatsApp, is likely to be dominant in the European Economic Area (“EEA”) market for consumer communication applications. The EC considers WhatsApp to be an important entry point to enable general-purpose AI assistants to reach consumers. In light of this, the EC’s provisional view is that the policy change is likely to be an abuse of this dominant position as Meta is refusing other businesses access to WhatsApp, including third-party AI assistants.
• Need for Interim Measures: The EC has stated that there is an urgent need for interim measures due to the risk of “serious and irreparable damage to competition”. Meta’s new policy risks raising barriers to entry and expansion, thereby marginalising smaller competitors on the general-purpose AI assistant market.

The EC’s investigation and Statement of Objections covers all of EU member states except Italy, to avoid overlap with the Italian Competition Authority’s ongoing proceedings (discussed in more detail below).  

The fact that the EC intends to impose interim measures – effectively a form of regulatory injunction – shows how seriously it is treating this case.  The EC faces a high substantive and procedural bar when seeking to impose such measures and will need to demonstrate that the immediate harm to competition caused by Meta’s AI policy justifies this form of emergency intervention.

Meta is likely to strongly contest the EC’s provisional findings, potentially citing some of the arguments outlined further below in this article.

It’s worth noting that Meta has already had ample opportunity to rehearse its arguments in the other global antitrust actions it has faced in relation to its new WhatsApp AI policy. These are summarised below.

Italian Interim Injunction (24 December 2025)

The Italian Competition Authority (ACGM) also launched a national antitrust investigation into Meta in mid-2025. In November 2025 it expanded this ongoing investigation to include the WhatsApp Business Terms and also launched proceedings for the adoption of interim measures to prevent Meta from making the change.

On 24 December 2025 the ACGM ordered Meta to halt its move to ban rival AI chatbots from WhatsApp’s Business platform in Italy pending the conclusion of their antitrust probe. In its 57-page decision, the ACGM warned that “... the sudden change in WhatsApp’s operating rules hinders and significantly changes [rivals’] development and investment plans, irreversibly impacting competition” and that “... such damage could prove disastrous for companies preparing to enter the market”. The Italian regulator publicly stated that Meta’s conduct appeared to constitute an abuse in the AI Chatbot services market, to the detriment of consumers and that they would be coordinating with the European Commission to ensure that Meta’s conduct is addressed in the most effective manner.

In response, Meta told the market that it would allow AI chatbots such as OpenAI, Poke.com and Luzia to continue to operate on WhatsApp in Italy while continuing to block the services outside Italy.

Brazilian Preventative Measure (12 January 2026)

On 12 January 2026 Brazil’s competition authority, the Administrative Council for Economic Defense (‘CADE’), ordered Meta to suspend its new contractual terms through a temporary injunction aimed at avoiding irreparable harm to competition and to ensure the effectiveness of its final decision. The order follows a complaint filed by Chatbot providers Luzia and Zapia who allege that the new WhatsApp Business Terms would unfairly prohibit millions of users from access to alternative AI services leaving Meta AI as the only option on the platform.

CADE found that there were strong indications of an abuse of dominant position by Meta under Brazil’s Competition Law and noted that their conduct could amount to an exclusionary strategy. The two key risks envisaged by CADE were that:

• Meta’s new policy could harm competition by excluding rivals and favouring their own assistant; and  
• The policy presented an imminent risk of irreversible damage to the market and consumers.

Meta was ordered to suspend enforcement of the new terms in Brazil and refrain from introducing similar restrictions until CADE has made its final ruling.

Initially, Meta complied with the order and adopted the same approach as in Italy, allowing AI providers to continue offering their chatbots to WhatsApp Business users with Brazilian phone numbers while continuing to block the services in the rest of the world.

However, on 23 January 2026 it was announced that Meta had won its appeal to suspend CADE’s injunction. The judicial case is under seal and Meta have stated that they “welcome this judicial decision to suspend the interim injunction from CADE. The facts do not support an intervention in Brazil or elsewhere” - no doubt a reference to the Italian interim injunction.  

Eastern and Southern Africa encouraged to follow suit  

On 5 January 2026, a further complaint was filed before the Common Market for Eastern and Southern Africa (‘COMESA’) Competition and Consumer Commission challenging Meta’s amendments to the WhatsApp Business Terms. The complaint is made under the COMESA Competition and Consumer Protection Regulations, 2025. It is alleged that Meta’s conduct raises serious concerns about restricting the cross-border provision of digital services within the Common Market and foreclosing competition in existing and emerging AI markets within the region.

As in other jurisdictions, the complainant in these proceedings is seeking interim action by the Commission on the basis that Meta’s conduct is capable of causing serious and irreparable harm to the competitive dynamics in the AI chatbot and assistant services markets within the COMESA region. As at the time of writing, no decision has been publicly reported yet.

What pro-competitive justifications has Meta raised?

Meta has consistently defended its new WhatsApp policy by arguing it was a necessary, pro-competitive measure rather than an attempt to stifle rivals. Meta has raised the following justifications in response to the allegations made in the proceedings described above:

• Protecting WhatsApp’s technical integrity and performance: Meta says third-party AI chatbots put undue strain on WhatsApp’s systems, which “were not designed to support” such high-volume, general-purpose AI interactions. This was a core point in Meta’s response to the European Commission and Italian proceedings - Meta called the Italian interim order “fundamentally flawed,” insisting the restrictions were needed to prevent technical overload and maintain quality of service
• Ensuring the platform is used as intended (preventing “unintended use”): Meta has characterised the presence of stand-alone AI chatbots on WhatsApp as an unintended misuse of its Business API. The company maintains that WhatsApp’s Business platform was built for facilitating direct business-to-customer communications not for independent AI services acting as third-party “apps” within WhatsApp. Meta argued in Brazil and other forums that the policy update was necessary to re-align WhatsApp’s use with its original purpose.
• Minimal impact on competition due to alternative channels: Meta’s public statements emphasise that the AI assistant market is highly competitive and that AI providers have many other ways to reach users. Meta has suggested that WhatsApp is not the primary distribution channel for AI chatbots so the policy does not foreclose competition. The company notes that people can access AI chat services via numerous alternatives – for example, through dedicated mobile apps, web platforms, search engines, email plugins, operating system assistants, or other messaging apps.  
• No blocker to AI innovation or consumer choice: Meta contends that its WhatsApp terms update will not harm innovation or consumer welfare because rival AI developers remain free to compete outside WhatsApp. Meta points out that its own AI assistant (“Meta AI”) competes in a crowded field of generative AI bots.  

Meta is now facing multiple investigations in different jurisdictions and it remains to be seen where competition regulators and the courts ultimately land on these issues. This may hinge on whether Meta can show that its policy is objectively justified and/or that it produces efficiencies that outweigh any anti-competitive effects.

Could the CMA take a similar approach in the UK?

There have been no public announcements from the CMA in relation to the updated WhatsApp Business Terms.

Some of the issues explored by the EC and other national competition authorities are consistent with the theories of harm identified by the CMA in its work on AI Foundation Models (see September 2023 initial report and subsequent update in April 2024). In particular the CMA’s Foundation Models (‘FM’) paper highlighted the risk that  “… powerful incumbents could exploit their positions in consumer or business facing markets to distort choice in FM services and restrict competition in FM deployment.”

That said, the CMA has not designated Meta / WhatsApp as having Strategic Market Status under the Digital Markets, Competition and Consumer Act 2024, which means the CMA would face the same balancing act in determining whether the WhatsApp policy is objectively justified or whether it constitutes an abuse of a dominant market position under Chapter II of the Competition Act 1998.

This publication is intended for general guidance and represents our understanding of the relevant law and practice as at February 2026. Specific advice should be sought for specific cases. For more information see our terms & conditions.