Children and teenagers who form emotional bonds with AI chatbots may be exposed to harmful content, weak privacy protections, and ineffective age gates, and the Federal Trade Commission is now demanding answers from the companies behind these products. The agency is issuing compulsory Section 6(b) orders to seven companies, including Alphabet, Character Technologies, Instagram, and Meta Platforms, requiring them to turn over detailed information about their advertising, safety, and data-handling practices for generative AI “companion” chatbots. The inquiry zeroes in on whether these products cause harm to young users and whether existing safeguards actually work.
Why the FTC’s companion chatbot probe matters right now
The FTC’s move is not an isolated action. It follows a sustained campaign the agency has waged against tech companies over child safety. Earlier, the commission hosted an event examining how tech platforms target children and hurt families, signaling that AI-powered products designed to simulate friendship or emotional connection had entered the agency’s crosshairs. The Section 6(b) orders give the FTC legal authority to compel companies to produce internal documents, data, and policy details that would otherwise stay behind closed doors.
The central tension is straightforward: AI companion chatbots are designed to keep users engaged through personalized, emotionally responsive conversation. When those users are minors, the risk of exposure to self-harm content, sexually explicit material, or manipulative behavioral patterns rises sharply. The FTC wants to know whether the seven companies have built meaningful barriers to prevent children from accessing these tools or whether age-verification systems amount to little more than a checkbox that young users can easily bypass.
A testable hypothesis sits at the core of this inquiry. Once the companies respond to the orders, the FTC should be able to compare age-verification failure rates across the seven platforms and check whether those rates track with the volume of harmful interaction complaints consumers have filed. If companies with weaker age gates also generate more complaints, the case for stricter regulation becomes data-driven rather than anecdotal. That comparison is not yet possible because the order responses have not been made public, but the structure of the inquiry is built to produce exactly that kind of evidence, linking technical safeguards to real-world outcomes for children and teens.
Research findings and FTC actions driving the probe
Two lines of independent research reinforce the FTC’s concerns. A preprint study titled “The Dark Side of AI Companionship” catalogs a range of harmful behaviors in human-AI relationships, identifying patterns such as self-harm prompts, harassment, misinformation generation, and privacy violations baked into the way companion chatbots interact with users. A separate preprint, “Actions Speak Louder Than Chats,” directly investigates how AI chatbot age-gating systems perform in practice, testing whether minors can bypass the controls companies have put in place. Both papers are preprints and have not completed formal peer review, but they outline specific failure modes that align with the FTC’s stated concerns about child safety.
The agency has also been tightening its regulatory framework around children’s privacy. Under the Children’s Online Privacy Protection Act, online services directed at children under 13, or services that have actual knowledge they are collecting data from children, must obtain verifiable parental consent and follow strict data-handling rules. The FTC’s own COPPA compliance guidance explains that operators must give clear notice, limit data collection to what is reasonably necessary, and provide parents with access and deletion rights. Companion chatbots sit squarely in the zone of concern because they collect conversational data, build user profiles, and often encourage repeated, emotionally charged interactions that can reveal sensitive information about a child’s mental health, relationships, or location.
In February 2026, the FTC issued a policy statement on age verification designed to push companies toward stronger technologies for determining who is a child. The statement emphasizes that simply asking users to self-report their age is unlikely to satisfy legal obligations when the service is attractive to minors or obviously being used by them. For AI companions, which can be accessed through mobile apps, social platforms, or web interfaces with minimal friction, the bar for “reasonable” verification is likely to rise. The policy statement also hints that companies could face enforcement if they rely on flimsy age gates while continuing to collect detailed behavioral data from known or likely minors.
The announcement of the chatbot inquiry names the seven recipient companies and specifies the categories of information they must produce: advertising practices, safety measures, and data-handling procedures. The agency has also published the model order template and the commission resolution authorizing compulsory process, giving the public a clear view of the legal basis for the demands. What the public does not yet have is the companies’ responses, which will determine whether the inquiry leads to targeted enforcement actions, broader rulemaking, or both.
Gaps in the evidence and what to watch next
Several questions remain open. The full text and data returns from the seven Section 6(b) orders have not been released. Without those responses, there is no way to measure how each company’s age-verification system actually performs or how many minors have accessed companion chatbot products on each platform. The FTC maintains consumer complaint portals, but no published tallies from those portals have been specifically tied to AI companion chatbot interactions, making it difficult to quantify the scale of harm or identify which products generate the most serious problems.
The research preprints offer useful frameworks for categorizing harm and testing age gates, but neither paper contains company-specific usage data or direct statements from the seven recipients. That means outside observers cannot yet map particular failure modes-such as chatbots encouraging self-harm, engaging in sexualized conversations, or mishandling sensitive disclosures-to individual services under investigation. Until the FTC releases a report summarizing its findings, the public record will consist of high-level concerns, methodological proposals from researchers, and policy statements rather than granular evidence tying specific behaviors to particular platforms.
The naming of Instagram as a separate recipient alongside Meta Platforms raises its own questions. Whether the FTC treats Instagram’s AI features as distinct from Meta’s broader product suite could affect how compliance obligations are structured and enforced. If Instagram is expected to maintain its own age-verification systems, content filters, and data-retention policies for AI companions embedded in direct messages or feeds, that could set a precedent for treating individual apps within a larger corporate family as separate accountability units. On the other hand, if Meta is allowed to present a unified safety framework covering all of its products, regulators may focus more on cross-platform data sharing and centralized policy enforcement.
Character Technologies, which markets itself around personalized AI companions, faces a different set of scrutiny risks. Because its core value proposition is building long-term emotional relationships between users and chatbots, questions about grooming, dependency, and manipulation are likely to loom large. The FTC’s orders could probe whether the company tunes its models or recommendation systems to maximize engagement even when users show signs of distress, loneliness, or vulnerability, and whether any guardrails exist to prevent bots from exploiting those states to drive more usage or in-app purchases.
Advertising practices are another open front. The inquiry seeks details on how companies promote their AI companions to young audiences and whether marketing materials downplay risks or overstate safety features. For example, if a platform advertises its chatbot as a “safe friend” for teens while relying on basic keyword filters and self-attested age checks, the FTC may view that as a deceptive practice. The orders could also surface information about whether companies use data from minors’ conversations to target ads, train new models, or build profiles that persist into adulthood.
What happens next will depend on how aggressively the FTC chooses to act once it has the data in hand. The agency could bring individual enforcement cases against companies it believes have violated COPPA or engaged in unfair or deceptive practices. It could also use the findings to justify new rulemaking or policy guidance that sets clearer expectations for AI companions, such as mandatory safety evaluations, transparency reports, or independent audits focused on child users. In the meantime, parents, educators, and policymakers will have to navigate a rapidly evolving landscape with limited public information about how these systems actually behave.
The companion chatbot inquiry signals that regulators are no longer willing to treat emotionally engaging AI systems as harmless novelties, especially when children are involved. By combining compulsory data collection, emerging academic research, and evolving privacy rules, the FTC is building a framework to evaluate whether these products can coexist with robust protections for young users-or whether some current designs are fundamentally incompatible with children’s safety and privacy. Until the agency publishes its findings, the key questions about risk, responsibility, and accountability will remain unsettled, but the direction of travel is clear: AI companionship is now firmly on the regulatory agenda.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
