Background
This ruling forms part of a wider piece of ongoing work on the advertising of AI products across a number of sectors.
The ads were identified for investigation following intelligence gathered by our Active Ad Monitoring system, which uses AI to proactively search for online ads that might break the rules.
See also related rulings published on 2 September, 16 September and 14 October 2026.
Ad description
Two paid-for Facebook ads by Antipadia Enterprises Ltd t/a Copymind: Ai twin for therapy, an AI therapy app, seen on 14 March 2026:
a. The first ad included text that stated, “Join thousands of users who have found clarity and direction, transforming stress into successful choices every day! [arm-muscle emoji]”.
The ad also included a video, split into three sections. The first section was a blurred, still image of a street view, with overlayed text that stated ““IF YOUR PAIN WAS EVER THIS BIG YOU WANTED TO FINALLY END THIS — TRY HEALING WITH YOUR AI TWIN”.
The central section depicted an emotional looking woman receiving a hug from a humanoid robot, and leaning into the robot’s chest.
The bottom section showed another still view looking down from the edge of a tall building, similar to the first section, but clearer, and with a person’s bare feet visible at the bottom of the image. The feet were positioned with the toes hanging over the ledge. Text over that image stated, “I DIDN’T CRY IN THERAPY. BUT I CRIED AFTER FIRST 20 MINUTES WITH MY AI TWIN. IT HELPED ME SEE PATTERNS I’VE BEEN STUCK IN MY WHOLE LIFE — WITHOUT JUDGMENT. AND FOR THE FIRST TIME, I STOPPED OVER-EXPLAINING AND JUST FELT SEEN.”
A voice-over read the text as the video was played.
b. The second ad showed a video of a woman doing daily activities such as using a laptop, looking out of a window and going for a walk, with calming music playing in the background.
A black phone screen-style box over the video contained the text “I SPENT THOUSANDS ON THERAPY AND STILL COULDN’T BE FULLY HONEST”.
“I was paying $150 an hour to… chat. We had no goals. No tools. No structure. Week after week, same conversations. She’d ask how my week was. We’d talk. She’d nod. Session over. And she never remembered anything from the week before. I realized [sic] I wasn’t making any progress. That’s why I created an AI copy of my brain in Copymind. I didn’t cry in therapy. But I cried after first 20 minutes with my AI Twin. It helped me see patterns I’ve been stuck in my whole life — without judgment. And for the first time, I stopped over-explaining and just felt seen. And it actually remembers everything “I don’t need someone to chat with once a week. I need daily support that helps me solve what’s stuck.”
A voice-over read the text as the video played but stated they had spent $200 an hour.
Issue
The ASA challenge whether:
- the ads made medical claims for unlicensed products;
- ad (a) discouraged essential treatment for conditions for which medical supervision should be sought; and
- the implied reference to suicide in ad (a) was likely to cause serious or widespread offence and cause unjustifiable distress, in particular to those affected by suicide and mental health conditions and was therefore irresponsible.
Response
Antipadia Enterprises Ltd t/a Copymind: Ai twin for therapy did not respond to the ASA’s enquires.
Assessment
The ASA was concerned by the lack of response and apparent disregard for the Code by Antipadia Enterprises Ltd t/a Copymind: Ai twin for therapy’, which was a breach of CAP Code (Edition 12) rule 1.7 (Unreasonable delay). We reminded them of their responsibility to provide a response to our enquiries and told them to do so in the future.
1. Upheld
The CAP Code stated that medicinal or medical claims and indications could only be made for a medicinal product that was licensed by the Medicines and Healthcare products Regulatory Agency (MHRA), the Veterinary Medicines Regulations (VMD) or under the auspices of the European Medicines Agency (EMA), or for a medical device with the applicable conformity marking. In addition, the Medical Devices Regulations required that a medical device should be registered with the MHRA before it was placed on the market in Great Britain.
In ad (a), the text stated, “Join thousands of users who have found clarity and direction, transforming stress into successful choices every day!”, “TRY HEALING WITH YOUR AI TWIN” and “I DIDN’T CRY IN THERAPY. BUT I CRIED AFTER FIRST 20 MINUTES WITH MY AI TWIN. IT HELPED ME SEE PATTERNS I’VE BEEN STUCK IN MY WHOLE LIFE — WITHOUT JUDGMENT. AND FOR THE FIRST TIME, I STOPPED OVER-EXPLAINING AND JUST FELT SEEN.” We considered those claims would be understood to mean the app could provide therapeutic psychological support, help users address long-standing emotional difficulties and produce a significant emotional breakthrough.
In ad (b), the text stated, “I SPENT THOUSANDS ON THERAPY AND STILL COULDN’T BE FULLY HONEST”, “I realized [sic] I wasn’t making any progress”, “I didn’t cry in therapy. But I cried after first 20 minutes with my AI Twin. It helped me see patterns I’ve been stuck in my whole life — without judgment. And for the first time, I stopped over-explaining and just felt seen”, and “I don’t need someone to chat with once a week. I need daily support that helps me solve what’s stuck.” We considered those claims would be understood to mean the app could provide more effective support than conventional therapy and help users resolve persistent and long-term mental health related difficulties.
We had regard to the MHRA’s guidance Digital mental health technology: qualification and classification, which set out what constituted a medical purpose in the context of digital mental health technology. We understood that claims to help treat or provide therapeutic support for mental health-related difficulties could amount to a medical purpose where they targeted symptoms or conditions that were diagnosable or clinically relevant.
We considered the claims in ads (a) and (b) went beyond general mental wellbeing. In particular, both ads referred to therapy and long-standing difficulties, while ad (a) referred to “healing” and ad (b) presented the app as providing ongoing support to help users “solve what’s stuck”. We considered those claims implied that the app could provide therapeutic support for clinically relevant mental health difficulties.
We sought advice from the MHRA, who said the ads made inferred medical claims and that the product was likely to fall within the scope of the medical device regulatory framework, which included software as a medical device.
We therefore considered the claims in both ads implied that the product had a medical purpose.
We had not seen evidence that the app was registered with the MHRA, and we understood the advertiser did not appear on the MHRA registration database. We had also not seen evidence that it held the applicable conformity marking. Therefore, no medical claims could be made for the product. Because the ads made such claims, we concluded that they breached the Code.
On that point, ads (a) and (b) breached CAP Code (Edition 12) rule 12.1 (Medicines, medical devices, health-related products and beauty products).
2. Upheld
The CAP Code stated that marketers must not discourage essential treatment for conditions for which medical supervision should be sought. For example, they must not offer specific advice on, diagnosis of or treatment for such conditions unless that advice, diagnosis or treatment was conducted under the supervision of a suitably qualified health professional.
We considered the claims in ad (a), “TRY HEALING WITH YOUR AI TWIN”, “transforming stress into successful choices every day”, and references to helping users “see patterns” they had been “stuck in” their “whole life” suggested the app could provide therapeutic support for serious mental health conditions. That impression was significantly strengthened by the ad’s imagery and wording which implied suicidal thoughts, including a still view from the edge of a tall building with a person’s bare feet visible at the ledge and the reference to wanting to “finally end this”. We considered consumers were likely to understand from that combination that the app could help people experiencing severe emotional distress, including conditions such as depression, for which medical supervision should be sought.
We had not seen evidence that any advice, diagnosis or treatment offered through the app was conducted under the supervision of a suitably qualified health professional. In the absence of such evidence, we concluded that ad (a) discouraged essential treatment for conditions for which medical supervision should be sought and therefore breached the Code.
On that point, ad (a) breached CAP Code (Edition 12) rule 12.2 (Medicines, medical devices, health-related products and beauty products).
3. Upheld
The CAP Code stated that marketing communications must be prepared with a sense of responsibility to consumers and to society. It also stated that marketing communications must not contain anything that was likely to cause serious or widespread offence and must not cause fear or distress without justifiable reason.
Ad (a) included the text “IF YOUR PAIN WAS EVER THIS BIG YOU WANTED TO FINALLY END THIS — TRY HEALING WITH YOUR AI TWIN” and imagery including a still view from the edge of a tall building with a person’s bare feet positioned over the ledge. We considered those elements together were a clear reference to suicide and that the combination of the imagery of someone contemplating jumping off a high building with the text “finally end this” was likely to cause distress to some people, particularly those who had been affected by suicide or were struggling with their own mental health.
We acknowledged that advertising which addressed mental health might include difficult material likely to cause distress. However, such scenarios should be treated with care must not cause fear or distress without justifiable reason. That might include, for example, raising awareness or directing people to support. If it could be justified, the fear or distress should not be excessive. However, as set out above, the ad was for an AI therapy app which we considered discouraged essential treatment and was not suitably licensed. In that context, we considered that the imagery and wording were not justified.
We also considered that the imagery and text in ad (a) was likely to cause serious offence for the same reasons.
For those reasons, we concluded that ad (a) was irresponsible, likely to cause unjustifiable distress and serious offence, and therefore breached the Code.
On that point, ads (a) breached CAP Code (Edition 12) rules 1.3 (Social responsibility), 4.1 and 4.2 (Harm and offence).
Action
The ads must not appear again in the form complained of. We told Antipadia Enterprises Ltd t/a Copymind: Ai twin for therapy not to make medical claims for their app unless it was registered with the MHRA and held the applicable conformity marking. We also told them not to discourage essential treatment for conditions for which medical supervision should be sought, and to ensure their ads were prepared with a sense of responsibility to consumers and to society and did not include references to suicide or imagery likely to cause serious offence or unjustifiable distress. We referred the matter to CAP’s Compliance team.
