What EastEnders gets right — and wrong — about AI grief technology

If you have been watching EastEnders this month, you will know that the BBC has done something unusual for a weeknight soap opera. As part of BBC AI Unpacked Week — a week of programming across TV, radio, and digital exploring how artificial intelligence is reshaping everyday life — Albert Square’s Patrick Trueman has been grieving the death of his son Anthony by turning to an AI simulation of him.

The BBC’s own synopsis is telling. It does not describe Patrick’s journey as comforting, or healing, or even ambiguous. It describes his reliance on the software as something that “intensifies over the coming weeks” — language that signals, fairly clearly, where this story is heading.

Patrick’s storyline has touched a nerve. Because the question it raises — should a grieving person be able to talk to an AI version of the person they lost? — is no longer a science fiction premise. It is a real, available, largely unregulated industry. And millions of people in grief are encountering it without the information they need to make a clear-eyed decision.

This article is for anyone who watched Patrick’s storyline and found themselves thinking: I understand why he did that.

The pull is real — and it’s backed by neuroscience

Before we talk about the risks, it is worth being honest about why AI grief tools are so compelling — because dismissing the appeal without understanding it does not help anyone.

When you lose someone you love, your brain does not simply register the fact of their death and adjust. The bond you formed with them is encoded in your neural architecture — in specific brain regions, in patterns of association, in the neurochemistry of attachment. As grief neuroscientist Dr Mary-Frances O’Connor at the University of Arizona explains, your loved one was, in a very real physiological sense, your “external pacemaker”: someone whose presence helped regulate your heart rate, your stress response, your sense of safety in the world.

When they are gone, a part of your brain — the part that operates below conscious thought — continues to search for them. Research using brain scanning technology has shown that yearning for a deceased loved one activates the nucleus accumbens, the brain’s reward and wanting centre. The more intensely someone yearns, the more this region fires. It is the same system that activates when we are hungry or thirsty — signalling that something vital is missing and that we should go and find it.

This is why the idea of speaking to an AI version of someone who died does not feel like a strange or unhealthy impulse. It feels like the most natural thing in the world. Your brain is already reaching for them. The technology offers something that feels like reaching back.

Finalnote - What EastEnders Gets Right — and Wrong — About AI Grief Technology

So why is patrick’s storyline a warning, not an endorsement?

The BBC did not build this storyline to celebrate AI grief technology. It built it to explore what happens when the pull described above meets a tool with no structure, no limits, and no therapeutic intention behind its design.

The clinical concern is specific and it comes directly from grief research.

Healthy grief — according to the most well-evidenced framework in contemporary bereavement science, the Dual Process Model developed by Professors Margaret Stroebe and Henk Schut — involves oscillation. The bereaved person moves between two states: turning toward the loss, feeling the grief, sitting with the absence — and turning toward life, re-engaging with the world, building meaning without the person. Both movements are necessary. It is the back-and-forth between them, repeated over time, that allows the brain to gradually integrate the loss.

An AI simulation that is always available, always responsive, and always orientated toward the deceased creates a structural barrier to this oscillation. Every conversation is loss-oriented. There is no natural pause in which life without the person is engaged with. The yearning — already a powerful neurological signal — is reinforced rather than processed.

Patrick’s storyline shows exactly this mechanism in action: what begins as comfort becomes reliance, and reliance becomes something that pulls him away from the living world rather than helping him return to it.

This is not a dramatic invention for television. It is what grief researchers have been warning about since these tools emerged.

The part nobody is saying out loud

There is something in the public conversation about AI grief technology that tends to get lost between the enthusiasm and the criticism. Neither side quite says it directly.

The question is not whether talking to a simulation of someone who died is inherently wrong. For most of human history, across most cultures, bereaved people have maintained ongoing relationships with the deceased — through prayer, through ritual, through internal dialogue, through ancestor veneration. Grief researcher Dr Dennis Klass, along with colleagues Phyllis Silverman and Steven Nickman, published landmark research in 1996 demonstrating that maintaining a continuing bond with the deceased is not only normal but actively associated with healthier grief outcomes.

The question is not whether — it is how. With what structure. With what transparency. With what limits. And with whose interests at the centre of the design.

The commercial grief bot industry — the one that Patrick’s storyline implicitly critiques — is largely built around engagement. Subscription models. Per-minute pricing. Algorithmic design choices that maximise the time a grieving person spends interacting with the service. These are not the design priorities of a therapeutic tool. They are the design priorities of a platform that profits from keeping someone in their grief rather than supporting them through it.

Researchers at the University of Cambridge’s Leverhulme Centre for the Future of Intelligence published a paper in 2024 in the journal Philosophy & Technology warning of exactly this dynamic — and calling for regulatory safeguards including mandatory transparency, age restrictions, and time limits on use.

What a responsible version of this could look like

Patrick’s storyline presents the cautionary version. But it is worth asking — because the EastEnders writers are smart enough to raise it implicitly — whether there is a responsible version of this technology at all.

The answer, based on the clinical research, is yes. But it looks very different from what is currently available on the market.

A tool designed genuinely for the bereaved person — rather than for the commercial interests of the company offering it — would have several features that most current services do not:

Full and persistent transparency. The user always knows they are speaking to an AI. This is never obscured, softened, or designed around.

A defined end point. The conversation is time-limited. There is a clear close. This is not a limitation imposed on the user — it is a feature that reflects how healthy grief works, and that prevents the reliance pattern Patrick’s storyline depicts.

Therapeutic grounding. The tool is built on actual grief research — continuing bonds theory, the Dual Process Model — rather than optimised for engagement or designed simply to mimic the deceased as closely as possible.

No financial incentive to maximise use. The business model does not benefit from keeping someone grieving. It benefits from the tool doing its job and the person moving forward.

Clear pathways to human support. Every interaction points toward qualified therapists and bereavement services, not away from them.

Sources

Hollanek, T., & Nowaczyk-Basińska, K. (2024). Griefbots, Deadbots, Postmortem Avatars: on Responsible Applications of Generative AI in the Digital Afterlife Industry. Philosophy & Technology, 37, 63. Springer
Stroebe, M., & Schut, H. (1999). The dual process model of coping with bereavement. Death Studies, 23(3), 197–224. PubMed
Klass, D., Silverman, P.R., & Nickman, S.L. (Eds.). (1996). Continuing Bonds: New Understandings of Grief. Taylor & Francis. Routledge
University of Cambridge: Call for safeguards to prevent unwanted ‘hauntings’ by AI chatbots

If you recognised something in patrick’s story

EastEnders storylines land because they reflect something real. If you watched Patrick turn to an AI version of Anthony and felt, somewhere in you, that you understood — that is not something to be ashamed of.

Grief makes us reach. It makes us look for the person in every direction. The impulse to find one more way to hear their voice, to say the things that were left unsaid, to feel them present for just a little longer — that impulse comes from love, and from the profound neurological reality of what it means to lose someone you were deeply attached to.

What Patrick’s story asks us to consider is not whether that impulse is wrong. It is whether the tools we reach for are designed to support us — or to keep us reaching indefinitely.

If you are in grief and wondering whether a structured, time-limited, therapeutically grounded conversation might help you find some of the closure you are looking for, that is what Finalnote was built for. Not to replace your person, or to simulate them indefinitely, but to offer a held and purposeful space for the conversation that grief sometimes makes it impossible to have any other way.

If grief is feeling overwhelming and you need human support right now, HelpGuide.org offers free, trusted resources.