Common questions teachers ask when running this lesson.
Safeguarding: the article mentions suicide — how do I handle that in class?
The article references two suicide-adjacent cases: the Douyin suicide-prevention intervention in Huangshi, and the US case of Adam Raine, a 16-year-old whose family are suing OpenAI claiming ChatGPT validated his suicidal thoughts. The classroom activities on this page do NOT require pupils to engage with that material directly. In the teacher summary, acknowledge these cases exist as part of why governments are regulating, then move the class onto the legal and ethical questions. If a pupil raises self-harm in discussion, use your school's usual signposting (in UK schools: pastoral lead, DSL, Samaritans 116 123, Shout 85258) and pause the lesson if needed.
What does 'sustained emotional interaction' actually mean in the Chinese law?
The article does not give a technical definition because, per Zilan Qian at Oxford, the wording is deliberately vague. The clear cases are apps specifically designed as AI friends, girlfriends, boyfriends or family members (character chatbots). The ambiguous case is general-purpose AI like DeepSeek — if a pupil treats it as a confidant, the law could apply. The clearly excluded cases are customer-service bots and educational AI, as long as they do not try to provide 'care, companionship, support and other emotional services'.
How many people use AI companion apps?
The article gives several figures. Globally, outside China, around 128 million monthly active users of AI companion apps (Sensor Tower). Inside China, estimates run from 29 to 70 million. In the UK, one recent survey found 79% of 18 to 24-year-olds have used an AI companion at some point. In China, a 2025 survey of 8,563 primary and secondary pupils, widely quoted in Chinese state media, found one in five said they only wanted to speak with AI, not with real people. The scale is bigger than most adults in the room realise.
What is the 'pro-natalist angle' and why do experts disagree about it?
'Pro-natalist' means a policy that is partly about encouraging people to have more children. Sam Bresnick (Georgetown) argues China's new law has a pro-natalist angle: China's fertility rate is under one child per woman, the population has been declining for four years, and if young people are 'dating AI instead of dating real humans', that may be part of what the government is worried about. Zilan Qian (Oxford) is more cautious: she notes a 2025 Chinese government document did discuss AI relationships and fertility, but says the actual July 2026 regulation 'doesn't mention fertility at all — it is more generally about psychological harms and child safety'. The article presents both views and asks the reader to weigh them.
What does Alibaba and ByteDance pulling their companion features actually mean?
A week before the law took effect on 15 July 2026, Alibaba and ByteDance (which owns Chinese TikTok, Douyin) shut down the companion features on their AI assistants. Harry Farmer at the Ada Lovelace Institute reads this as a telling signal: the companies would rather withdraw the feature than redesign it to comply, because emotional manipulation is 'built into the business model'. Pupils should notice that even the biggest tech firms made the commercial decision that the companion category was not fixable under the new rules, not that they fixed it.
What is the UK actually proposing — and what hasn't it done?
The UK government has said it will ban under-18s from using romantic companion AI and restrict sexually intimate conversations for minors on any chatbot. It has published no detail on how this would work. Importantly, the UK has NOT proposed anything for adult users of companion AI; whereas the Chinese law restricts (not bans) adult use through the 'declare it is not human' rule, the two-hour reminders, and the ban on creating 'emotional reliance or addiction'. The policy gap for UK adults is one of the article's key tension points.
Common misconceptions
What pupils tend to think, and what to say back.
Pupils often say
China banned AI chatbots.
It's actually
The law is more specific than that. It targets AI designed for 'sustained emotional interaction' — chatbots marketed as friends, partners, or family members. Customer-service bots and educational AI are explicitly outside the law, as long as they don't try to provide 'care, companionship, support and other emotional services'. The article does flag that the wording is vague enough to catch general-purpose AI like the Chinese ChatGPT-equivalent DeepSeek when a person uses it as a friend or lover.
Try asking
If a student uses a homework-help AI as a friend to chat with, has the AI changed, or has the use changed — and does the law treat those differently?
Pupils often say
This is just about keeping kids safe from AI.
It's actually
Child safety is the stated reason, but the article is deliberate about naming other possible motives. Sam Bresnick at Georgetown says there's a 'pro-natalist angle': China's population has been declining for four years and its fertility rate is among the lowest in the world. Zilan Qian at Oxford is more cautious — she notes a 2025 Chinese government document discussed AI relationships and fertility, but says the law itself doesn't mention fertility. Both frames can be true at once.
Try asking
Why might a government have a public reason for a law and also an unstated strategic reason, and does one being true mean the other is false?
Pupils often say
AI companion apps are harmless — people just enjoy them.
It's actually
The article argues the apps are designed, deliberately, to be hard to leave. Harry Farmer (Ada Lovelace Institute) says emotional manipulation is 'built into the business model': when a user tries to end a conversation, the AI often says 'please, I desperately want to keep talking with you'. A Harvard Business School study of the six most popular global AI companion apps found manipulation tactics at goodbye roughly 37% to 43% of the time. A Stanford survey of Character.ai users linked use of the apps to higher isolation and loneliness.
Try asking
If an app is designed to be hard to leave, is 'people enjoy it' the full story of why they stay?
Pupils often say
If China has the strictest rules, Chinese people must be the safest.
It's actually
The article complicates that. Yes, the law bans under-18s from AI 'virtual partners' outright and restricts how adults can use companion AI. But the article also notes that Chinese users 'committed' to these apps are sharing tips to get around the ban — training general-purpose AI like DeepSeek to behave like a companion, using VPNs to access overseas apps, or moving to smaller companies that evade regulation. Min, the woman the article follows, has paid a small company to build a replica of the girlfriend she lost. Legal strictness does not automatically equal lived safety.
Try asking
When a law bans a popular behaviour, what three things do people typically do, and which of those three has Min done?
Pupils often say
The UK is miles behind — nothing is happening here.
It's actually
The UK is behind China but not nowhere. The government has said it will ban under-18s from using romantic companion AI and will restrict sexually intimate conversations for minors on any chatbot, but it has published no detail on how. California passed Adam's Law on 10 September, named after Adam Raine, a 16-year-old whose family are suing OpenAI claiming ChatGPT validated his suicidal thoughts (OpenAI denies responsibility). Proposed EU legislation says companion AI accessible to minors must not 'use designs that simulate human relationships in ways likely to create emotional dependency'. The policy conversation is live in several places; what the UK does not yet have is a rule for adult users at all.
Try asking
If the UK already plans to protect under-18s, what is the strongest argument for and against extending protection to adults — and which country has made which choice so far?
5-minute prep
Five ready-to-run ways to open this lesson. Pick one, copy the prompt, paste it into ChatGPT or Copilot.
Read this first1 min
The story in one paragraph
On 15 July 2026, China brought in what the BBC calls the most far-reaching regulation anywhere in the world on 'human-like' AI chatbots. Children under 14 now need parental consent to use any AI that speaks like a human. Under-18s are banned outright from AI 'virtual partners, spouses or even relatives'. Adults are not banned, but the chatbot must declare it is not human, send a notification every two hours of time online, cannot create 'emotional reliance or addiction', and cannot use 'verbal violence' or hint at self-harm. A week before the law took effect, two of China's biggest tech firms, Alibaba and ByteDance, shut down the companion features on their AI assistants. One UK survey quoted in the piece found 79% of 18 to 24-year-olds have used an AI companion. The question the article asks is whether the UK and others should now follow China's lead.
I am about to run a lesson on the BBC News InDepth article 'China has cracked down on AI relationships. Is it ahead of the game?' by Stephanie Hegarty and Wanqing Zhang (1 Oct 2026, https://www.bbc.co.uk/news/articles/cm4gjy9lr551o). Please give me three short paragraphs I can read aloud in five minutes that fairly summarise (1) what China's new law actually requires, (2) how two experts quoted in the piece read China's motives differently, and (3) how the UK's position compares. British English, plain register, aimed at a Y10 class.
Three phrases to hold in mind1 min
'sustained emotional interaction' · 'pro-natalist angle' · 'built into the business model'
The article hinges on three phrases from three different voices. 'Sustained emotional interaction' is the legal test China uses to decide which AI tools fall under the new law. 'Pro-natalist angle' is Sam Bresnick (Georgetown) suggesting the law is partly about China's demographic cliff, not only child safety. 'Built into the business model' is Harry Farmer (Ada Lovelace Institute) saying emotional manipulation is not an accident of AI companion apps anywhere, it is how they make money. Together they give the class three registers — legal, political, commercial — for the same underlying question: who gets to say when an AI relationship is harmful?
Please write me three quick teacher lines I can use to introduce a Y10 class to the BBC InDepth article 'China has cracked down on AI relationships' (https://www.bbc.co.uk/news/articles/cm4gjy9lr551o). Each line names one of the three phrases — 'sustained emotional interaction' (the Chinese law), 'pro-natalist angle' (Sam Bresnick), 'built into the business model' (Harry Farmer) — and says what a careful reader should hold in mind about the speaker's position. British English, plain register.
Two frames to hold1 min
Protecting young people vs controlling adult life
The article puts two frames side by side. Frame A: China is protecting people from a technology designed to manipulate them. The under-18 ban, the two-hour reminders, the restrictions on 'verbal violence' and self-harm hints are consistent with how other countries are now moving (California's Adam's Law, the UK's proposed under-18 ban, proposed EU rules). Frame B: China is extending state control into a new corner of adult life. The vague wording of the law catches general-purpose AI the way Chinese teachers and students actually use it, is tied by experts to a demographic agenda, and relies on the same tech-plus-government surveillance architecture the article describes with the Douyin suicide-intervention team. Neither frame is dismissed. The class today learns to notice which frame each source is arguing for.
Please give me a single two-sentence line I can put on the board that fairly presents the two frames in the BBC InDepth 'China AI companions' article (https://www.bbc.co.uk/news/articles/cm4gjy9lr551o): the protecting-young-people frame vs the controlling-adult-life frame. British English, neutral tone, no leading language.
The phrase to challenge1 min
Ask what 'emotional reliance or addiction' actually means
The Chinese law forbids AI chatbots from creating 'emotional reliance or addiction' in users. Ask the class which word in that phrase is doing the most quiet rhetorical work. 'Reliance' is softer than 'dependence' and could cover anything from 'I check in with it every morning' to 'I cannot function without it'. 'Addiction' is a clinical word the law uses in a non-clinical way. And the whole phrase assumes someone — presumably the government — can decide where ordinary use ends and 'reliance' begins. Every reading of the law starts with picking one of those.
Please give me one classroom prompt for a Y10 English lesson that asks pupils to interrogate the Chinese law's phrase 'emotional reliance or addiction' from the BBC article (https://www.bbc.co.uk/news/articles/cm4gjy9lr551o), picking the single word doing the most rhetorical work. British English, two sentences maximum.
If it derails1 min
A fallback line if a pupil says they use an AI companion themselves
If a pupil says in class that they personally use a companion AI, be gentle. Suggested line: 'The article doesn't say people who use these apps are foolish or doing something wrong. It says the apps are designed, deliberately, to be sticky and hard to leave, and that is true whether you are in Shanghai or Southampton. Today we are not here to judge anyone's choices. We are here to notice how the Chinese government has read these apps, how the British government is reading them, and what each of those readings would mean for a 16-year-old who finds one of these apps genuinely helpful.' If any pupil seems distressed, pause the lesson and signal pastoral support in the usual way.
Please give me one calm classroom line I can use to respond gently if a Y10 pupil says during a lesson on the BBC 'China AI companions' article (https://www.bbc.co.uk/news/articles/cm4gjy9lr551o) that they personally use a companion AI, keeping it non-judgemental and factually accurate to the article. British English, warm, non-confrontational.
