The decision in brief
According to publicly reported information, the Competition Board (Rekabet Kurulu) decided on 14 May 2026, in decision no. 26-18/536, to open a full investigation into Meta Platforms, Inc., Meta Platforms Ireland Limited, WhatsApp LLC and Meta's Istanbul-based entity. The investigation concerns whether Meta abused a dominant position, contrary to Article 6 of Law No. 4054 on the Protection of Competition, by changing WhatsApp's business terms in a way that excludes third-party general-purpose AI assistants.
Alongside the investigation, the Board imposed interim measures under Article 9 of the same law. In essence, Meta must ensure that providers of third-party general-purpose generative AI chatbots and assistants can access WhatsApp without being hindered, whether de facto or through economic conditions, while the investigation continues. Meta was given one month from notification of the reasoned decision to comply. The decision became public in early June 2026 through reports in the Turkish press.
Interim measures do not mean the Board has found an infringement. They are a temporary tool designed to prevent harm that may be difficult to reverse by the time a full investigation is concluded. But their very adoption signals that the Board sees a serious prima facie concern.
Background: how WhatsApp became an AI distribution channel
With billions of users worldwide and near-universal penetration in Turkey, WhatsApp is not simply a messaging app: it is the place where people already spend their digital day. For AI companies, being reachable in a WhatsApp chat removes the friction of downloading a new app or visiting a website. Over 2024 and 2025, several AI assistants — including well-known chatbots from major developers — launched WhatsApp numbers through which users could talk to them directly.
These services relied on the WhatsApp Business Solution, the interface Meta offers companies for communicating with customers at scale. In October 2025, Meta announced changes to the Business Solution terms, effective 15 January 2026, prohibiting providers of AI technologies from using the interface to distribute general-purpose AI assistants where the AI is the primary functionality being offered. Businesses using AI for incidental purposes — such as a retailer's customer-service bot — remained permitted. Several AI providers subsequently announced they would withdraw their assistants from WhatsApp.
At the same time, Meta has been integrating its own assistant, Meta AI, directly into WhatsApp. Critics argued that the combination — excluding rivals from the channel while placing Meta's own assistant inside it — is a classic pattern of leveraging a dominant platform into an adjacent, fast-growing market.
Turkey is not alone: the European front
The Turkish decision is part of a wider regulatory response. In Italy, the competition authority (AGCM) opened an investigation in 2025 into Meta's integration of Meta AI into WhatsApp, extended it to the new Business Solution terms, and in December 2025 adopted interim measures ordering Meta to suspend the terms' exclusion of rival AI chatbots for Italian users. In the same month, the European Commission opened a formal antitrust investigation into whether the new policy breaches EU competition rules.
These parallel proceedings matter for two reasons. First, they show that authorities with different procedures reached a similar provisional concern independently. Second, they raise the prospect of a patchwork: a global platform may face different access obligations in different countries while investigations run their course. Meta has publicly disputed the concerns, arguing among other things that the Business Solution was not designed to carry general-purpose chatbots and that such traffic strains its systems.
The legal framework: Article 6 and Article 9 of Law No. 4054
Abuse of dominance
Article 6 of Law No. 4054 prohibits undertakings holding a dominant position from abusing it. The provision offers a non-exhaustive list of abusive conduct, including preventing others from entering the market or hindering competitors' activities, and making the sale of one product conditional on another. Dominance itself is not unlawful; the question is whether conduct distorts competition on the merits.
In a case like this, the Board would typically examine several questions:
- Market definition. Is the relevant market consumer messaging services in Turkey, business messaging services, or a distinct market for access to messaging users? The narrower and more "must-have" the channel, the stronger the dominance finding.
- Dominance. Market shares, user penetration, network effects, switching costs and multi-homing behaviour all matter. Messaging apps benefit from powerful direct network effects: people use the app their contacts use.
- Abuse. Excluding a category of business users from an essential channel, while the platform's own competing service remains inside it, may be analysed as a refusal to supply, discriminatory treatment, or self-preferencing.
- Objective justification. Meta may argue that the restriction protects system integrity, user experience, security or privacy. Any justification must be genuine, necessary and proportionate — and less restrictive alternatives will be scrutinised.
Interim measures
Article 9 of the Law empowers the Board to take interim measures during an investigation where there is a likelihood of serious and irreparable harm. In digital markets, the concern is "tipping": once users form habits around one AI assistant inside their messaging app, rivals may be unable to recover their reach even if the conduct is later found unlawful. Interim measures aim to freeze the competitive landscape until the Board reaches a final decision.
Consequences of non-compliance
Law No. 4054 provides for administrative fines if interim measures are not complied with, including periodic fines calculated for each day of non-compliance on the basis of the undertaking's turnover. If the Board ultimately finds an infringement, it may also impose fines of up to ten per cent of annual turnover and order behavioural or structural remedies. The Law's commitment and settlement mechanisms, introduced in 2020, allow the investigated party to offer commitments or settle — options that may shape how the case unfolds.
Turkey's track record against Meta and big tech
The decision fits a pattern. In January 2021, the Board adopted interim measures against WhatsApp's controversial privacy-policy update, which would have expanded data sharing with Facebook, ordering that users in Turkey not be bound by the new data-sharing terms during the investigation. In 2024, it imposed interim measures concerning the combination of data between Instagram and Threads, prompting Meta to suspend Threads in Turkey temporarily. The Board has also issued a series of significant decisions against Google on self-preferencing in search, shopping and local services.
Taken together, these cases show a competition authority that is willing to use interim measures quickly and to treat data, defaults and access to user bases as core competition issues. We have also examined the Authority's attention to algorithmic conduct in our report on its dynamic-pricing investigation.
Why distribution is the new battleground in AI
The WhatsApp case highlights a shift in how competition in AI is understood. Much early debate focused on compute, chips and training data. But as capable models become more widely available, distribution — access to users at the moment they need help — is becoming the decisive bottleneck. Operating systems, browsers, app stores, productivity suites and messaging apps are the gateways.
Whoever controls those gateways can shape which assistant users meet first, which becomes the default, and which gathers the interaction data that improves future models. That dynamic has already produced tensions between platform owners and AI developers; we discussed one example in our report on the legal tension between Apple and OpenAI. As assistants evolve into agents that take actions on users' behalf — a trend we explored in our analysis of an autonomous agent's misconduct — control over the channels they operate through will matter even more.
Regulation by design: the Digital Markets Act and Turkey's draft
In the EU, the Digital Markets Act (DMA) addresses gatekeeper conduct ex ante, without the need to prove dominance and abuse case by case. WhatsApp is designated as a core platform service, and the DMA already imposes interoperability obligations on gatekeeper messaging services. Those obligations concern messaging between users of different services, however, and do not squarely resolve access for AI chatbots — which is one reason authorities have turned to traditional competition tools.
Turkey has worked for several years on amendments to Law No. 4054 that would introduce DMA-style obligations for undertakings with "significant market power" in core digital services. The draft has not yet been enacted. The WhatsApp case illustrates both the value of such rules and the capacity of the existing Article 6–Article 9 toolkit to act in the meantime.
The personal-data dimension
Opening WhatsApp to third-party AI assistants raises privacy questions that competition law alone does not answer. When a user talks to a third-party chatbot on WhatsApp, messages flow through Meta's infrastructure to the AI provider, which then processes them — often on servers abroad. Under the KVKK, AI providers serving users in Turkey must identify their legal basis, inform users, secure the data and comply with cross-border transfer rules. Users often share sensitive information with assistants, including health and financial details that may fall into special categories.
Platforms, for their part, may invoke privacy and security as justifications for restricting access. Competition authorities increasingly take such arguments seriously but examine whether they are pretexts. The Board's 2021 WhatsApp decision itself shows how data protection and competition analysis can intertwine in Turkey. For the compliance framework applicable to AI providers, see our guide to KVKK and AI governance.
What the decision means for different players
For AI developers and start-ups
The interim measures preserve a valuable distribution channel for AI assistants in Turkey, including Turkish start-ups building local-language assistants. But access comes with responsibilities: robust privacy notices, security controls, abuse prevention and clear information for users that they are talking to an AI system. Providers should also be ready for Meta to set technical and commercial conditions — which must not, under the interim order, amount to de facto exclusion.
For platforms
The case is a reminder that platform terms of service are not merely contractual documents; they can be competition-law conduct. Changes that disadvantage rivals in adjacent markets — especially where the platform offers a competing service — should be reviewed for competition risk before launch, with a documented, evidence-based justification and consideration of less restrictive options.
For businesses using WhatsApp
Companies that rely on the Business Solution for customer service are not the target of the restriction, but they should monitor how terms evolve and ensure their own AI-powered bots comply with transparency and data-protection obligations.
For users
Users in Turkey may continue to reach a choice of AI assistants inside WhatsApp during the investigation. That choice is also a responsibility: before sharing sensitive information with any assistant, users should check who operates it and how their data will be used.
Compliance checklist: platform terms as competition risk
- Map adjacent markets. Identify where your platform's business users compete with your own products or services.
- Assess dominance honestly. Consider network effects, user penetration and switching costs in each national market, not only global shares.
- Document justifications. Record the technical, security or quality evidence behind any access restriction, and why narrower measures would not work.
- Apply rules consistently. Treat your own services and rivals' services under the same conditions, or be ready to explain objectively why not.
- Plan for divergence. Build the ability to apply different access regimes by jurisdiction while investigations proceed.
- Coordinate with privacy teams. Ensure data-protection arguments are genuine and consistent with how the platform treats its own services.
How the investigation could end: four scenarios
A full investigation under Law No. 4054 typically takes many months and includes written defences, an investigation report and an oral hearing. Its outcome could take one of several shapes:
- Closure without infringement. If Meta persuades the Board that the restriction was objectively justified and proportionate, or that WhatsApp is not dominant in a properly defined market, the case could be closed and the interim measures lifted.
- Commitments. Since the 2020 amendments, undertakings under investigation can offer commitments to resolve competition concerns. Meta could, for example, commit to transparent and non-discriminatory access conditions for AI providers. If the Board accepts binding commitments, the investigation can end without a fine.
- Settlement. The investigated party may also acknowledge the infringement and settle, in exchange for a reduction in the fine. This path is less likely where a company disputes the core theory of harm, but it remains available.
- Infringement decision. If the Board finds an abuse, it may impose a fine calculated on turnover and order Meta to end the conduct, which would in practice turn the interim access obligation into a lasting remedy. Such a decision would be subject to judicial review.
The scenario that materialises will also depend on developments in Europe: a commitment package accepted by the European Commission, for instance, could become a reference point for the Turkish authority.
Why this matters for Turkish-language AI
The case has a less visible but important local dimension. Large global assistants can reach Turkish users through many channels. Smaller Turkish developers building assistants tuned for Turkish language, law, public services or local commerce have far fewer options. For them, being reachable in the messaging app that nearly every smartphone user in Turkey already uses can be the difference between viability and obscurity. An AI ecosystem where only the platform owner's assistant is present in the everyday channel would make it much harder for local innovation — one of the aims of the 2026–2030 AI Action Plan — to reach users at scale.
What to watch next
Several milestones will shape the case. First, how Meta implements the interim measures in Turkey, and whether the technical and commercial access conditions it sets are accepted by the Board as non-exclusionary. Second, whether Meta challenges the interim measures before the administrative courts. Third, whether commitments are offered — for example, transparent, non-discriminatory access terms for AI providers. Fourth, the direction of the parallel proceedings in Italy and at the European Commission, which may influence the substance of any final remedy. Finally, whether Turkey's draft digital-markets amendments gain new momentum.
Frequently asked questions
Has the Competition Board found that Meta broke the law?
No. The Board opened an investigation and imposed interim measures. A finding of infringement, if any, will come only at the end of the investigation after Meta has exercised its rights of defence.
What exactly must Meta do?
According to the reported decision, Meta must ensure that third-party general-purpose generative AI chatbots and assistants can access WhatsApp without de facto or economic hindrance during the investigation, and was given one month from notification of the reasoned decision to comply.
Does this affect customer-service bots used by businesses?
The restriction at issue concerned general-purpose AI assistants. Businesses using AI incidentally for customer service were not excluded under Meta's terms, so the interim measures mainly protect general-purpose assistant providers.
What happens if Meta does not comply?
Law No. 4054 allows the Board to impose administrative fines for non-compliance with interim measures, including periodic fines for each day of non-compliance calculated on the basis of turnover.
Can Meta challenge the decision?
Yes. Competition Board decisions, including interim measures, can be challenged before the administrative courts in Ankara. A challenge does not automatically suspend the obligation to comply.
Are other authorities looking at the same issue?
Yes. Italy's competition authority adopted interim measures in December 2025, and the European Commission opened a formal investigation in the same month.
Does opening access create privacy risks?
It can. Messages sent to third-party assistants are processed by those providers, who must comply with the KVKK, including transparency, security and cross-border transfer rules. Users should be careful about sharing sensitive information.
Expert Opinion
This section reflects my personal assessment as the founder of this site and an AI ethics & compliance counsel.
In my view, the Competition Board's decision is one of the most consequential AI-governance moves Turkey has made this year, even though it does not mention "AI regulation" at all. It recognises a simple truth: in the AI economy, the most valuable asset may not be the best model but the shortest path to the user. When a platform that already sits at the centre of people's daily communication closes that path to rivals while opening it to its own assistant, competition law has good reason to step in early. Waiting years for a final decision could mean ruling on a market that has already tipped.
That said, interim measures are a blunt instrument, and the Board will need to handle the next steps carefully. Meta's arguments about system integrity and user experience deserve a genuine hearing, and "access" must be defined in a way that is technically workable and does not create new privacy or security risks. The best outcome would be transparent, non-discriminatory access conditions that apply equally to Meta AI and to third-party assistants — ideally shaped through commitments rather than years of litigation.
My advice to AI companies is to treat this decision as an opportunity, not a guarantee: build distribution strategies that do not depend on a single gatekeeper, and make privacy compliance a selling point rather than an afterthought. My advice to platforms is to treat every change to access terms as a potential competition-law event. And for Turkish policymakers, this case is a strong argument for finalising a clear, predictable digital-markets framework — so that fairness in AI distribution does not depend on case-by-case interventions alone.
This article is for information only and does not constitute legal advice. Facts about the decision are based on publicly reported information, including press reports published in June 2026; the reasoned decision may contain further details. The analysis and assessments are the author's own.
