AI companions, dating apps and the privacy of digital intimacy

21 November 2026 · By Sarah Newbold, therapist at Progressive Therapeutic Collective

What feels private when you type it? A question about sex. A message to someone you fancy. A fear about your dad that you can barely admit to yourself. In the moment, your attention is on what you need to say and who might answer. The company holding the conversation can feel a long way away.

Six stories from this week brought those two experiences closer together. The first begins with a daughter trying to make sense of her father’s dying.

1. When a daughter turned to ChatGPT during her father’s final illness

Journalist Bridget Todd was already investigating AI companionship when her father became seriously ill. In a WBUR interview aired on 20 November, she described turning to ChatGPT for support. She’d explained in an earlier interview how those conversations changed. During the day, she had medical questions. At night, she found herself going back over decisions about her father’s care, wondering whether a different choice might have saved him. Listen to the WBUR interview.

What stayed with me was how familiar that question is. Did I do enough? can be an extraordinarily hard thing to ask someone who loved him too. If someone told me they’d taken that question to a chatbot, I’d want to know what the reply gave them, and whether they’d like someone in their life to hear the question as well.

2. What happens to private chats when an AI companion company goes bankrupt?

On 20 November, Pennsylvania’s attorney general announced that the US bankruptcy court overseeing Character.AI’s parent company had ordered the appointment of a consumer privacy ombudsman. The independent expert will examine any proposed sale involving users’ accounts or chat data, including children’s conversations, and recommend safeguards to the court. The judge will have that advice before deciding whether a sale can proceed. Read the attorney general’s announcement.

Think about a teenager trying out the words I think I might be queer for the first time. They might choose a fictional character precisely because telling someone at school feels too exposing. They deserve to know, in language they can understand before they type, who keeps those words and what could happen to them if the company changes hands.

3. What can Grindr Edge tell someone about you?

On 20 November, a Grindr user described their surprise at seeing figures on other profiles about how often people replied or started conversations. It was one person’s account of encountering Edge, Grindr’s paid AI features, which launched worldwide on 17 November after early testing in Australia and New Zealand. Read the user’s account.

Grindr says that, with your consent, it can use your profile, activity and chats to infer patterns such as the age groups you tend to contact. Subscribers may see insights about you even if you use the free app. There’s a difference between telling a potential date what you like and having an app draw its own conclusions from people you’ve messaged. You can see and delete individual insights under Privacy Settings → Insights. To switch off this personalisation, go to Privacy Settings → Sensitive Data → Personalize gAI Features. Read Grindr’s guide to Edge and privacy settings.

4. Sexual deepfakes and consent in Jess Asato’s case against Grok’s operator

British MP Jess Asato is suing SpaceXSI after users created and shared sexualised images of her using Grok. The company says it removed material that breached its terms. For some images it left online, its lawyers argue that viewers would recognise them as fictional and that Asato therefore had no reasonable expectation of privacy in them. Asato is asking the High Court to require technical measures that would stop Grok generating manipulated images of her. The court has yet to rule on those arguments. Read Reuters’ account of the claim and defence.

A colleague might know perfectly well that an image is fabricated. You still have to walk into work knowing they’ve seen your likeness used sexually. Whether anyone was fooled is one question. What it costs a person to have other people decide how their body is depicted is another.

5. Why a company is selling AI girlfriends to executives worried about spies

On 20 November, WIRED reported that some matchmakers working with men in tech had heard an unusual dating worry. Clients feared that someone who seemed romantically interested might be after company secrets. One business spotted an opening and began marketing AI girlfriends to executives on the promise of protection from romantic espionage. Read the WIRED report.

I’m interested in what that sales pitch asks a buyer to take on trust. On a date, you can say, “I’d rather keep work out of this,” and learn something from how the other person responds. With a commercial AI companion, the reassuring reply comes from a character. The decision about what happens to anything you disclose belongs to the company behind it.

6. Deepfake calls and romance scams: what to do when someone asks for money

On 20 November, The Wall Street Journal reported on criminals using AI-generated voices and video to impersonate executives and persuade employees to transfer money. In an October public awareness campaign, INTERPOL described how the same technology can exploit trust between family members or people getting to know each other online. Read the overnight reporting and INTERPOL’s guidance.

If someone you love calls in distress, of course you want to help. A sentence worth having ready is “I’m here. I’m going to ring you straight back.” End the call and use the number you already have for them. If someone you’ve only met online asks for money, pause the transfer and talk it through with someone you trust. If you’ve already sent money, contact your bank straight away.

Next
Next

Who Checks the AI That Screens for Family Violence Risk?