Digital Legacy and the Afterlife of Data: What Happens to You Online When You’re Gone

The Immortality You Never Asked For

Nobody sits down to plan their digital estate. You plan your will, maybe — if you are the organized type, or if someone in your family scared you with a story about a distant cousin and a disputed lakehouse. You plan your funeral arrangements, possibly. You might even have an awkward conversation with your family about what to do with the record collection. But your Google account? Your twelve years of Instagram posts? The 47,000 text messages on your phone, including the ones you definitely should have deleted? These things have a way of slipping through the cracks of mortal planning.

And yet here is the reality: digital data is remarkably persistent. The servers that host your accounts are not sentimental, but they are sturdy. Your Facebook profile, your email archives, your cloud storage full of blurry photos from weddings you barely remember attending — all of this will outlast you, absent deliberate action to the contrary. We are the first generation of human beings to face this reality in any meaningful way, and most of us are handling it by not thinking about it at all.

This is worth thinking about. Not in a morbid way — or at least, not only in a morbid way — but because the questions it raises are genuinely profound: Who owns a dead person's data? What happens to the people left behind who want access to it? Can a digital presence constitute a form of legacy, the same way a published book or a painted portrait does? And perhaps most unsettling of all: should we be building AI systems that can simulate dead people, and if so, what does that do to grief, to memory, and to the nature of identity itself?

The Legal Tangle

Let us start with the practical, because the practical turns out to be a mess. When you sign up for any major platform — Apple, Google, Facebook, Amazon — you agree to terms of service that typically state, somewhere in the fine print that exactly no one reads, that your account is personal and non-transferable. This means that, legally speaking, when you die, your account does not pass to your heirs the way your furniture does. It reverts, in a sense, to the platform.

What this looks like in practice has generated some genuinely heartbreaking legal cases. There was the family of a German teenager killed in a train accident who spent years in court fighting Facebook for access to her account — they wanted to read her final messages to understand whether her death was an accident or a suicide. Facebook initially refused on grounds of privacy. A German court eventually ordered access in 2018, but the legal process took five years and required the family to fight all the way to the Federal Court of Justice. Five years. For access to their own daughter's thoughts.

Apple has similarly resisted providing access to deceased users' accounts, and for years had no official policy for handling such requests at all. It took a grieving father in Arizona — who had to obtain a court order just to access his late wife's iPad photos — before Apple finally introduced a Legacy Contact feature in 2021, allowing users to designate someone who can access their account after death. It was a straightforward, obviously necessary feature. It took a lawsuit and a decade of people dying with locked devices to get it.

The legal landscape is improving, slowly and unevenly. The Revised Uniform Fiduciary Access to Digital Assets Act, adopted in varying forms by most U.S. states, creates a framework for digital estate planning. The European Union's General Data Protection Regulation takes a different approach, treating the data of deceased persons as generally outside its protections — leaving the question to member states, who have handled it inconsistently. The upshot is that if you die without having made explicit provisions for your digital accounts, the process of accessing or inheriting them can be legally complex, emotionally exhausting, and, in some cases, simply impossible.

The Platforms Have the Power

The uncomfortable reality at the center of all of this is that the platforms hold the power. They wrote the terms. They control the servers. They decide, ultimately, whether a grieving family gets access or gets a politely worded denial. This is not merely an abstract legal observation. It means that some of the most intimate records of a human life — conversations, photographs, journals written in the form of social media posts — are held in private custody, subject to corporate decisions that may have nothing to do with the wishes of the person who created them or the needs of the people who loved them.

The Memorialization Economy

Platforms have responded to the reality of user death in a variety of ways, some more thoughtful than others. Facebook introduced memorialized accounts, in which a deceased user's profile is preserved and accessible to friends and family for remembrance but the account can no longer be logged into. A designated Legacy Contact can manage the account — posting remembrances, responding to friend requests — but cannot read private messages or change significant account settings.

Google has an Inactive Account Manager, which allows users to set up an automated plan: after a defined period of inactivity, designated contacts can be notified and given access to specified data. Twitter, now X, will deactivate an account upon proof of death if requested by a verified family member, but does not provide access to private data. Instagram follows a similar memorialization model to Facebook.

What is interesting about these policies is not only what they allow but what they reveal about how platforms think about death. They have built systems that preserve the appearance of a person — the profile, the posts, the public-facing archive — while carefully maintaining control over the infrastructure beneath it. A memorialized Facebook account is not really an estate. It is a display case. The actual data — the messages, the behavioral patterns, the vast machine-learning-friendly archive of preferences and interactions — remains with the company, quietly.

The Business of the Dead

And this raises a question that does not get asked often enough: what are companies doing with the data of dead users? Legally, in most jurisdictions, the answer is: whatever they were doing with it before, unless specific deletion is requested and honored. The terms of service do not typically include a "death clause" that halts all data processing. In a world where data is routinely used to train AI systems, to refine algorithms, and to build behavioral models, the data of deceased users is potentially just as valuable as the data of living ones — and considerably less likely to complain about how it is being used.

The AI Resurrection Question

This is where we enter territory that is genuinely new, genuinely contested, and — depending on your philosophical disposition — either deeply comforting or profoundly disturbing. Over the past several years, a small but growing industry has emerged around what might be called digital resurrection: the use of AI to create interactive simulations of deceased individuals, based on their existing data.

The most prominent example is the company HereAfter AI, which offers a service allowing people to record their stories and memories during their lifetime, which are then made available to family members as an interactive AI after their death. A related company, StoryFile, worked with the actor William Shatner to create an interactive AI version of himself — a project that raised immediate questions about what would happen to that simulation after Shatner himself died.

More controversially, the startup Eternos — and others operating in the same space — offer services that can generate a simulation of a deceased person using their existing digital footprint: their social media posts, their emails, their texts, their voice recordings. Using large language models and voice synthesis, these services claim to produce an entity that speaks, responds, and even jokes in a manner consistent with the person it is based on.

The Ethics of Speaking for the Dead

The ethical questions here are significant and genuinely unresolved. Does a person have the right to consent — or to refuse consent — to being digitally resurrected? Most people have never thought about this question, which means the default answer is usually determined by whoever holds the data and whoever holds the grief. If your family wants to create an AI version of you and they have access to your digital footprint, there may be nothing in current law to stop them, regardless of what you would have wanted.

There is also the question of grief itself. Does interacting with an AI simulation of a deceased loved one help the grieving process or distort it? The psychology research on this is nascent, but early indicators are cautious. Some grief counselors worry that these simulations may complicate the acceptance stage of mourning by providing a substitute for the person rather than a path toward living without them. Others argue that any technology that reduces suffering is worth using, and that humans have always found ways to maintain symbolic relationships with the dead — through photographs, through rituals, through the stories we tell about them.

What seems clear is that AI resurrection raises questions that photographs and rituals do not: the simulation can respond. It can generate new content. It can, in a meaningful sense, participate in a conversation. Whether what it produces has anything to do with what the actual person would have said is a question that neither the technology nor its proponents tend to answer with much directness.

Digital Legacy as Cultural Memory

Stepping back from the individual and the legal, there is a larger cultural dimension to all of this that deserves attention. The aggregate of our digital data — the vast, sprawling, searchable archive of what hundreds of millions of people posted, searched, shared, and said across a generation — constitutes something genuinely unprecedented in the history of human civilization. No previous era has left anything like this record.

Historians of future centuries will have access to the inner lives, the daily rhythms, the opinions and anxieties and jokes of ordinary people at a level of granularity that is simply not available for any previous period. The diary of a medieval peasant is an almost unimaginable treasure. In two hundred years, the equivalent of ten billion simultaneous diaries will be available, if anyone can figure out how to access and interpret them.

This is, depending on your perspective, either a remarkable gift to the future or an extraordinary privacy violation in both directions — committed by the present against the future, and by future generations against the present. The people who posted their most intimate thoughts on early-2000s LiveJournal blogs almost certainly did not contemplate the possibility that digital archivists would still be mining those posts for cultural insight in the twenty-third century.

What You Can Actually Do

Given all of this, the practical question becomes: what should you do? The answer, frustratingly, is that most people do nothing, and the gap between what is legally possible and what people actually arrange is enormous.

At minimum, the experts in digital estate planning recommend: creating a document that lists all your digital accounts and passwords, stored somewhere secure and accessible to a trusted person; designating Legacy Contacts or equivalent features on platforms that offer them; deciding explicitly what you want to happen to each major account — preserved, deleted, or transferred; and incorporating your digital estate into any formal will or estate plan with the help of a lawyer who knows what they are doing in this area.

Beyond the practical, there is a philosophical decision each of us faces: what kind of digital presence do we want to leave behind, and for whom? The answer to that question is not a legal or technical matter. It is a deeply personal one, and it is one that most of us are currently leaving to chance — or to the terms of service.

LET'S GET CRITICAL

The article you just read makes a strong case for taking digital legacy seriously — the legal tangle, the ethical questions around AI resurrection, the unprecedented cultural archive we are building. All of that is real and important. But there are several threads in this conversation that deserve to be pulled harder, and some assumptions in the mainstream digital legacy discourse that are worth questioning.

Start with the framing of digital legacy as primarily a problem of access and preservation. The implicit assumption running through most discussions of this topic — including, to some extent, the article above — is that more access to a deceased person's data is generally better than less. Families should be able to access their loved one's accounts. Digital estates should be inheritable. The record should be preserved. But this framing takes as given something that is actually quite contested: the idea that a person's digital footprint is a meaningful representation of who they were.

Consider what your digital data actually captures. It captures what you searched for, what you posted publicly, what you messaged to other people in what you believed was a private context. It does not capture the texture of your daily thinking, the things you considered and decided not to say, the kindnesses that left no digital trace, the person you were in a room when no devices were present. A family that gains access to a deceased person's complete digital archive may find that what they receive is simultaneously overwhelming in volume and surprisingly thin in the things that actually mattered. More data is not necessarily more understanding.

This connects to a deeper problem with the AI resurrection discussion. The article presents digital resurrection as an ethical gray area, which it is. But one thing that does not get said clearly enough is that these systems are, structurally, impressive hallucinations. A large language model trained on your text data will generate outputs that sound like you, but the mechanism by which it does this is pattern matching and statistical prediction — not memory, not understanding, not anything resembling the actual cognitive and emotional processes that made you who you were. When a grief-stricken parent talks to an AI trained on their child's texts, what they are talking to is not a version of their child. They are talking to a very sophisticated mirror of their own expectations, shaped by training data that the child produced in a completely different context and with no anticipation of this use.

The danger here is not just philosophical. There is genuine psychological risk in systems that blur the distinction between simulation and person. Our capacity for grief depends, at some level, on the reality of loss — on the recognition that the person is gone and will not return. A technology that creates a persistent, responsive simulacrum that can be interacted with indefinitely is not a tool for processing grief. It may be a tool for indefinitely deferring it. The grief counselors who have raised this concern are not being sentimental or technophobic. They are applying professional knowledge about how human minds process loss, and their caution deserves more airtime than it typically gets.

There is also a power and equity dimension that the article touches on but does not fully develop. The companies building digital legacy tools — the HereAfter AIs and Eternoses of the world — are not charities. They are commercial entities operating in a largely unregulated space, building businesses around one of the most emotionally vulnerable markets imaginable: people in grief, people facing their own mortality, people who would give a great deal to have one more conversation with someone they have lost. The potential for exploitation in this market is enormous, and the current regulatory framework offers almost no protection.

The data sovereignty question also runs deeper than the article's framing suggests. When we talk about families wanting access to their loved ones' accounts, we tend to discuss it as though the only parties involved are the family and the platform. But those accounts contain conversations with other living people — people who may have shared things in what they believed was a private exchange with the person who is now dead, and who have a legitimate privacy interest in those communications. The right of a family to read their deceased daughter's messages is not simply the family versus the platform. It is also the family versus every person who sent that daughter a private message and never consented to a third party reading it.

Finally, and perhaps most importantly, we need to be much more honest about the class and cultural dimensions of digital legacy planning. The advice to "consult a digital estate lawyer" and "create a comprehensive digital asset inventory" is sound, but it describes a practice that is accessible primarily to people with the time, education, and resources to engage with it. The people most likely to die young — due to poverty-related health outcomes, violence, or occupational hazard — are also the people least likely to have arranged their digital estates. Their families may be the ones who most acutely need access to a deceased person's device or account and the ones with the fewest legal resources to pursue it. A conversation about digital legacy that does not grapple with this disparity is, at minimum, incomplete.

FANTASTIC GUEST | A Conversation with Jorge Luis Borges