Part of the AI Regulation News hub.
DSIT's government response to the Growing up in the online world consultation commits to mandatory usage breaks on AI chatbots for every child under 18
The headline in June was a social media ban for under-16s. The July response goes after something narrower and harder to design around: how long a child is allowed to keep talking to a chatbot without being interrupted.
Bottom line: A policy document, not law. The response states commitments; the department says detailed definitions of which chatbots fall in scope will be set out in regulations in due course.
Who this affects: Product and safety leads at consumer AI chatbot and companion-app developers, online safety counsel, child-safety officers at gaming and social platforms, and clinical governance staff at digital mental health services offering conversational tools.
Issue date: Government response published July 2026, added to the consultation page on 15 July 2026. The underlying consultation ran from 2 March 2026 to 26 May 2026.
What changed: In June the commitment covered sexualised chatbot content. July adds mandatory breaks for all under-18s, an explicit threat to ban children from certain chatbots over mental health advice, and an exemption for business and customer service chatbots.
Analysis: The exemption line matters more than the ban line. A general assistant that a teenager uses for homework is not obviously a customer service tool, so the scope question falls entirely to regulations that do not exist yet.
Primary sources: Consultation page and outcome documents · Government response, July 2026 (HTML) · Government response, web-accessible PDF, 44 pages
- Instrument (EN)
- Growing up in the online world: government response (July 2026)
- Authority
- Department for Science, Innovation and Technology (DSIT)
- Jurisdiction
- United Kingdom
- Status
- Consultation outcome published; implementing regulations not yet made
- Bindingness
- Not binding in itself. Commitments are to be delivered through regulations and regulator action
- Reference
- CP 1643, ISBN 978-1-5286-6739-5, E03646827 07/26, 44 pages
- Issue date / next deadline
- July 2026, published to GOV.UK on 15 July 2026 / no compliance date stated
- Stated legal vehicle
- Regulations, not yet made; the response names powers under the Children's Wellbeing and Schools Act 2026 only for later changes needed to keep protections effective
- Primary source
- https://www.gov.uk/government/consultations/growing-up-in-the-online-world-a-national-consultation/outcome/growing-up-in-the-online-world-government-response-july-2026
The commitment, in the department's own words
Chapter 3 of the response deals with chatbots directly. The department writes that it will require AI chatbots to introduce mandatory breaks for all children under 18, and says it will work with experts on how frequent and how long those breaks should be.
That expert work is routed through an Expert Panel chaired by Professor Russell Viner, which the response describes as supporting government to assess the scientific evidence related to digital safety.
No frequency, duration or start date appears anywhere in the chapter. The commitment is to the mechanism, with the parameters deferred.
Why breaks, rather than an age gate
The response gives its evidence base openly. Over half of young people aged 16 to 21 who responded, 59 percent, wanted restrictions on chatbots that encouraged longer conversations. Eighty percent of respondents to the full consultation, and 68 percent of 16 to 21 year olds in the children's consultation, thought there were particular risks where chatbots mimic friendship.
On mental health, 18 percent of 16 to 21 year olds responding to the children's consultation indicated that young people see chatbots as a source of emotional or mental health support.
The department is candid that it is balancing two things. It quotes the Centre for Information Policy Leadership on the value of purpose-built educational chatbots and the Royal College of Psychiatrists on the risk of systems that appear authoritative, relational, anthropomorphic or emotionally responsive.
The second commitment carries a threat
Alongside breaks, the response says the government will work with regulators to take action on chatbots that share harmful mental health advice, and will consider all options, including banning children from certain chatbots where other regulatory and safety mechanisms are not enough.
That is conditional language, not a ban. It also names the machinery: work with the Department of Health and Social Care, NHS England, and the Medicines and Healthcare products Regulatory Agency, which the response notes can act where AI products within its remit pose risks to public health, including chatbots that claim to diagnose, treat or manage a mental health condition.
The MHRA is described as reviewing how AI in healthcare should be regulated, with a response to recommendations due later this year. Anyone building a conversational wellbeing feature should be tracking that review rather than this document.
The scope gap the response admits to
The department states that the majority of chatbots most used by children are already within scope of the Online Safety Act, which requires action on illegal content and on content harmful to children. It then says some chatbots are not covered by the Act because of the way they function, and that the government will act swiftly to close that loophole.
It also carves out: chatbots that are typically used in business or customer service settings will be exempt from the measures, and detailed definitions of which chatbots are in scope will be set out in regulations in due course.
Between an admitted gap in the Act and an undefined exemption, the practical scope of the break requirement is unsettled. A product team cannot currently read this document and know whether its assistant is caught.
Where this sits in the wider package
The chatbot measures arrive attached to much larger commitments: banning social media companies from providing services to under-16s, preventing under-16s from accessing harmful functions on other services including gaming, and default restrictions for 16 and 17 year olds covering an overnight window between midnight and 6am, muted overnight notifications, and autoplay and personalised feeds off by default.
The June announcement, dated 15 June 2026 in the Secretary of State's foreword, already covered sexualised chatbot content: preventing under-18s from accessing AI chatbot services that primarily offer sexualised content, and barring other chatbot services from offering children features enabling sexually explicit or sexual role-playing content.
The response states that if changes are needed to keep protections effective, the government will use powers under the Children's Wellbeing and Schools Act 2026 to make them.
What we did not verify
We opened the GOV.UK consultation page and the HTML government response of July 2026, and read the chatbot chapter, the summary, the foreword and the implementation passages in full.
We did not open the 44-page print-ready PDF, the June 2026 progress statement, the children and young people's summary, the summary of evidence and methodology, or the Welsh translations. We did not read the Children's Wellbeing and Schools Act 2026 or the Online Safety Act 2023.
We make no claim that mandatory breaks are in force, that any specific product is in scope, or that a break of any particular length has been decided. The response states none of those things.
If you ship a conversational product that under-18s can reach, the actionable item is not the break itself but the scope question. The department has deferred the definition of an in-scope chatbot to regulations and has exempted business and customer service tools without saying where the line falls. Document now how your product is used by minors, because that record is what a scope argument will rest on later.
Source File
Open the July 2026 government response and confirm three passages: section 3.1 stating that chatbots will have to introduce regular breaks for children under 18, the sentence committing to consider all options including banning children from certain chatbots, and the line stating that chatbots typically used in business or customer service settings will be exempt with definitions to follow in regulations.
We will therefore require AI chatbots to introduce mandatory breaks for all children under-18. ยท DSIT, Growing up in the online world: government response, CP 1643, July 2026
FAQ
Is the mandatory break requirement law now?
No. The response is a policy document. It says the requirement will be delivered through regulations, and it does not state a commencement date or a break length.
Which chatbots are covered?
Not yet settled. The response exempts chatbots typically used in business or customer service settings and says detailed definitions of in-scope services will be set out in regulations in due course.
What did the government already commit to in June 2026?
Preventing under-18s from accessing AI chatbot services that primarily offer sexualised content, and preventing other chatbot services from offering children features that enable sexually explicit or sexual role-playing content.
Does this affect mental health chatbots specifically?
It signals action. The response commits to working with regulators on chatbots that share harmful mental health advice and notes the MHRA can act on chatbots claiming to diagnose, treat or manage a mental health condition.
Related briefings
Sponsored Training
Practical AI training for regulated professionals, built around verification, documentation and a defensible process. See the courses.