About This Research

Why is this important, and why now?

Who is this research for?

This research is intended for technical founders, startup executives, and decision makers for investment into early-stage companies focused on the digitizing of human essence, and who wish to connect their work to current digital ethical scholarship.

What is important to know?

The digitization of human essence in an era of artificial intelligence goes way beyond what we’ve experienced with social media. It is the deliberate, conscious effort to capture as much about a living human prior to death. With this information, those who survive can posthumously interact with the departed’s essence in a number of different chatbot-based executions. From simple text responses predicated on mapping best-fit responses to an existing database of recorded answers, to the simulation of live video and the believable generative simulation of a person interacting from beyond the grave via a VR headset. Such approaches counteract traditional notions of grief, transcend our understanding of what it means to be alive, and for many who’ve already done this, feel as if they put a stopper in death.

Why now?

We’re increasingly accustomed to depictions of de-aging in the movies, but what of the essence capturing and resurrection of loved ones, or even ourselves? What responsibilities do developers shoulder in building these products? We already store large volumes of information about ourselves in our computers’ memories, but artificial intelligence is now able to remix such storage with our actual memories in simulating human interaction. We know it’s not real, but it feels real. It is … enough.

Legislation around such practice of harvesting is already years behind, and the pace of generative platforms is only increasing as services such as ChatGPT accelerate in scale and capability. This research is intended as a timely intervention and reminder of some of the ethical considerations for those invested in the success of these products prior to their scaling to mass audiences. As custodians of human life and those accountable for its preservation in perpetuity, ingraining the human dimensions of care and empathy are deeply important for the future of these platforms. Through injecting ethical considerations and tactical examples early into a large language model’s training, we can reduce bias, discrimination, and issues of ownership, privacy and identity which inevitably arise from such work.

What are the ethical issues this research seeks to unpack?

These ethical issues include, but are not limited to:

Deeply important issues of privacy, permissions and the commercialization of digital identity are highly applicable to those who cannot speak for themselves. This might include the posthumous use of digital likeness, but also of digital essence through chatbots and voice replicator services.

The proactive extraction of an individual’s persona where permission is given prior to death, as in the case of Hereafter Ai. What are the tactical methods used, what are the implications for issues of faith, and how do they impact a person’s digital agency and ability to control their own narrative after they have passed? What are the ethical implications of interacting with a digital essence for those left behind, and for whom the product is often intended?
How do these technologies reframe our sense of self as individuals, and what permissions are we giving those who seek to monetize our digital identities? What do we understanding about the value exchange left behind for those who survive us as the end users?
Unintended consequence and the true cost of such a value exchange. How might this technology be used for sinister ends, despite the best of intentions of its creators? This is where I intended to not just follow the money, but follow the intended outcomes of such potentially darker uses.

Why is this important to me?

Later this year I will turn 50. And the older I get, the more I’ve begun to write down the stories of my life for my daughter to read one day. The stories of how I met her mother. The stories of surviving cancer. The stories of my move to America. These stories don’t seek to just capture the events of what happened, but the feelings at the time. What it felt like to see my wife for the first time. What it felt like to have the doctor give me the all-clear. What it felt like to walk through a departure gate to my new life. The stories I’ve written down capture the essence of a life lived. Each piece always starts the same:

While I write literally all day at work - emails, slack messages, whiteboards and more, I haven’t actually sat down and truly written anything in many years. Encouraged by my wife, Mary, to start sharing some of the stories of my life, especially with our daughter, Emma, I’m going to take on the challenge of writing down as much as I can remember about the moments which have been important to me along the way. Emma, I hope one day you’ll read and enjoy these as much as I’ve enjoyed remembering them.

The products of grieftech and the harvesting of digital essence as one ages or declines is a natural place I’ve arrived at, aged 49. But compounded with the increasingly dystopian risks of generative artificial intelligence, how might we take the emotionally positive sentiments of a written memory and digitize it in a more ethically-influenced way?

What are the next steps for this research?

My hope is that the thoughts explored in this research might serve as a place to pause before launch for products intersecting with these ideas of digital essence. That they provide a place to think about not just what’s next, but what’s best for most without further marginalizing the already marginalized. That they offer a toolkit for how to infuse early-stage products with the ability to scale ethically, rather than react to problems when it’s already too late. I hope they offer actionable thoughts on the difference between can and should. And I hope they serve as an ethical blueprint for organizations seeking to genuinely offer comfort to the bereaved.