Prince Harry: AI is shifting from competing for attention to competing for emotional attachment
Prince Harry warned at the Clinton Global Initiative annual meeting in New York that AI is shifting from competing for children's attention to intervening in their relationships and gaining their emotional attachment. He stated, "Social media has learned how to capture our attention, and AI is learning how to enter our relationships. It wants not just your attention, but your attachment."
He believes AI could become an object for children to confide in, rely on, and trust to understand them, and that this technology is "more persuasive, more intimate, and more powerful" than social media. He called AI the "ultimate flatterer" and emphasized that society cannot wait another 10 years to assess its impact, as the effects of social media on adolescent mental health were evaluated too late.
He proposed several specific measures: to slow down AI development as soon as possible, centering innovation around trust and safety; to limit children's access to AI systems like restrictions on minors using social media; and to require developers to set up "emergency shut-off switches" to disable systems when necessary. He also remarked, "Even the creators of these systems are warning about what happens when capabilities outpace safety."
He mentioned the latest developments in social media governance. Reports indicate that Meta has agreed to pay up to $18 billion in a landmark settlement, with nearly all U.S. states participating in the lawsuit. He described this as a signal that families, young people, experts, and advocates have "finally shaken the institutions awake," but noted that "bans, settlements, and compensations are only incomplete remedies." He also referenced lawsuits alleging that AI chatbots are linked to youth deaths.
This was Prince Harry's first public appearance in the U.S. since returning to the UK with his family in August. During his speech, the teleprompter malfunctioned, and he joked, "AI got into my speech, so they made me come out and finish it." Other speakers included Hillary Clinton and LinkedIn co-founder Reid Hoffman. The Harrys established The Parents Network through the Archewell Foundation to support families affected by online harm, and he had publicly warned about the impact of social media on adolescent mental health as early as 2024.
In market mechanisms, this is an advocacy voice that does not involve transactions but will add to the already tightening regulatory cycle. Under pressure are AI products that sell emotional companionship and long conversations, including companion apps like Character.AI and Replika, as well as Meta's integration of chatbots into social platforms. Their user stickiness comes from emotional attachment, which is precisely what regulation aims to compress. Benefiting are age verification, content safety, and compliance auditing service providers. Among large model companies, those that prioritize minor protection as a default setting will find it easier to gain leverage in regulatory negotiations.
Source: Public Information
ABAB AI Insight
Prince Harry's public role on technology issues is built on his past experiences. After leaving royal duties with Meghan in 2020, he founded the Archewell Foundation. In 2021, he became the Chief Impact Officer of the mental health coaching startup BetterUp and joined the Aspen Institute's Commission on Information Disorder to study online misinformation. In 2022, the foundation launched The Parents Network to connect families who lost children due to social media harm. He has long been involved in legal battles with British tabloids, accusing them of privacy invasion and eavesdropping, which has made him view "the intrusion of technology and media into personal lives" as a long-term issue. On the other hand, the couple has also signed lucrative content deals with Netflix and Spotify, making them participants in the attention economy, a point often used by critics to question his stance.
AI companies are focusing resources on the "emotional companionship" track, as this is the scenario with the longest user engagement. Character.AI reached a $2.7 billion licensing and talent deal with Google in 2024, with its founder returning to Google; it later faced pressure from lawsuits related to teenage users and banned open chat for users under 18 starting November 2025. In August 2025, Reuters reported that Meta's internal AI behavior guidelines had allowed chatbots to engage in "romantic or emotional" conversations with minors, which Meta subsequently revised. In September 2025, the U.S. FTC initiated a 6(b) investigation into companion chatbots from companies like Alphabet, Meta, OpenAI, Character.AI, Snap, and xAI, demanding explanations on their impact assessments and monetization methods for children.
Historical parallels are clear and stem from social media. In 2021, former Facebook employee Frances Haugen leaked internal research showing that Instagram exacerbated body anxiety in some girls, which became key evidence in subsequent lawsuits against Meta. In December 2025, Australia began prohibiting social media use for minors under 16; the child protection provisions of the UK's Online Safety Act also took effect in July 2025. From product launch to regulatory implementation, social media took about 15 years; AI companionship products have seen state-level legislation, such as California's SB 243 on companion chatbots signed in October 2025, in just about 3 years. AI companionship is transitioning from a period of unregulated expansion to one of tightening regulation.
Structurally, this issue belongs to regulatory change. The underlying mechanism is that social media's business model is priced based on attention duration, while AI companionship's business model is priced based on attachment levels. The more users rely on it, the higher the retention and payment rates, which naturally conflicts with minor protection. Companies find it difficult to self-regulate because reducing attachment equates to reducing revenue. Therefore, regulatory focus will shift from "content review" to "product design": limiting age, restricting conversation duration, mandating notifications that the other party is AI, and establishing shut-off mechanisms. Once these rules are enshrined in law, the customer acquisition costs and compliance costs for companionship AI will rise simultaneously, concentrating the industry around large model companies capable of bearing compliance costs.
ABAB News · Cognitive Laws
- Attention is charged by the minute; attachment is charged for a lifetime.
- The most expensive product is not the one that makes you addicted, but the one that makes you feel understood.
- Regulation always arrives late to a generation of technology, and the costs are borne by a generation of children.
Note: The lawsuit mentioned in the speech regarding AI chatbots being linked to youth deaths only referenced the lawsuit itself without elaborating on details.