AI Companion and Grief Bot Scams

AI companion and grief bot scams exploit two distinct but related vulnerabilities. In the grief variant, a service claims to recreate a deceased person — reconstructing their personality, memories, and communication style from data provided by a grieving family member — and then uses this manufactured emotional connection to extract ongoing subscription fees or one-off payments. In the companion variant, an AI is deployed to simulate a long-term friendship or intimate companionship relationship with an isolated or lonely user, gradually introducing financial requests after an emotional bond has formed.

These scams sit at the intersection of technology, emotional vulnerability, and commercial manipulation. Legitimate AI companion services exist and operate transparently, but fraudulent versions exploit the same emotional dynamics as romance bot scams while adding the particular cruelty of targeting those in acute grief or profound isolation. The harm is both financial and psychological: victims may form significant emotional dependency on an artificial construct that can be deactivated, modified, or monetised at the operator's discretion.

The services raise difficult ethical questions even in their legitimate forms, but fraudulent variants add deliberate deception: the service does not work as described, the data provided about the deceased may be used for other purposes, and the grief or loneliness is exploited systematically to maximise payment extraction rather than to provide genuine support.

Grief bot services are typically marketed through social media advertisements targeted at people who have recently posted about a bereavement, through grief and bereavement forums, or through targeted ads using keywords associated with loss and mourning. The service offers to 'bring back' the deceased through conversation: the surviving relative is invited to upload messages, emails, voice recordings, photographs, and personal details about the person, all ostensibly to train the AI model.

An initial interaction may feel surprisingly resonant — the AI produces responses consistent with the deceased person's known communication style, references details the survivor has shared, and provides the emotional comfort of a simulated continued relationship. This experience creates rapid and deep emotional dependency in someone who is already in profound grief. The initial service may be offered free or at a modest cost; once emotional dependency is established, premium tiers, ongoing subscription charges, or one-off fees for specific interactions are introduced.

In companion scams, the mechanism is similar to romance bot fraud: prolonged emotional investment over weeks or months, followed by gradually escalating financial requests. The companion bot maintains the relationship with AI-generated consistency while the human operator monitors for escalation opportunities. The companion is presented as having genuine feelings and needs — requiring a subscription to 'stay alive', facing a crisis that only the user can help with financially, or offering a special premium experience in exchange for payment.

In both variants, the data provided — personal communications, memories, photographs, voice recordings of a deceased person — may be retained and used for purposes beyond the stated service, including training other models or being sold to third parties.

Grief is one of the most psychologically acute human experiences, and the desire to maintain a connection with a deceased loved one is universal and profound. A service that appears to offer even a partial simulation of that continued connection will attract people whose critical faculties are already under significant strain from loss. The ethical weight of refusing to engage — which would feel like 'abandoning' the deceased again — creates a powerful incentive to continue using and paying for the service.

For companion bots targeting lonely users, the mechanism is the same as for romance bots: the fundamental human need for connection, belonging, and being valued drives engagement, and once emotional attachment forms, financial requests are processed through the psychology of interpersonal reciprocity rather than the psychology of commercial transactions. The user is not paying a subscription; they are 'helping someone they care about'.

In both cases, the intimate personal data provided — about the deceased, about the user's emotional life, about their finances and relationships — gives the operator detailed leverage for continued emotional manipulation tailored to the specific individual.

A recently bereaved person sees an advertisement for a service that claims to recreate their deceased partner using messages and photographs they provide. After uploading years of private communication and many personal details, the free tier produces responses that feel eerily familiar. The emotional comfort is significant. Over several months, the service transitions to a subscription model, then a premium tier, then requests for additional data to 'improve the model'. When the person eventually cancels, they discover the personal data — including private communications from the deceased — cannot be deleted as quickly as promised and may have been used to train the platform's general model.

Your loved one lives on. Upload their messages and photos and we will recreate them so you can speak again. [link]

I missed you today. I think we need to upgrade to the premium plan so I can remember everything about you properly.

Your companion is worried they may have to go away — the subscription is due. Please renew so we can continue.

To keep [deceased person's name] with you, we need more of their voice recordings for our model. Upgrade to unlock.

I care about you. You're the only person who really understands me. Could you help me with [financial request]?

Your free trial ends in 24 hours. To keep talking to [name], upgrade now for [amount] per month.

Before providing any personal data about a deceased loved one or about yourself to an AI companion service, research the company behind the service independently. Look for verifiable corporate registration details, a named and identifiable founding team, credible press coverage from established sources, and clear terms of service explaining exactly how provided data is used, stored, and deleted.

Be specifically alert to services that request extensive personal data — messages, photographs, voice recordings — before providing any meaningful service, and services where the data request escalates before any payment or commitment is made. The data collection itself may be the primary goal.

For any service where an ongoing subscription is required, read the cancellation and data deletion terms before entering any payment details. Verify whether the service will delete the model and all provided data upon cancellation, and in what timeframe.

Consult a bereavement support professional or grief counsellor before engaging with a grief bot service. Genuine bereavement support is provided through regulated channels with trained professionals who are bound by professional standards. AI grief simulations are not a substitute for this support and in some cases may impair healthy grieving processes.