For the foreseeable future, and probably forever until the heat death of the universe, this will always be true: Everyone dies. But it’s still not an easy thing for many of us to accept, as the rising belief in transhumanism illustrates. Rest assured, one day the light will go out, whether we want to think about it or not.
Increasingly, it’s looking like chatbots are going to be the ones helping some of us to accept the inevitable and wrestle with our very human questions about death. Researchers are already making bots that can talk people through their personal and, importantly, spiritual anxieties about death in addition to helping them plan for the end. What that says about us, I’m really not sure.
End-of-life care planning isn’t an easy thing to do, but it’s harder in some places than in others. In the US, for example, it wasn’t until last year that Medicare, which covers people over the age of 65, covered end-of-life consultations. As soon as it was covered, more than double the expected number of seniors took advantage. But that policy only helps some people — tens of millions of Americans are currently uninsured.
Chatbots like LifeFolder’s Emily have popped up to mimic conversations with healthcare workers to help people plan. Emily, according to the company, is meant to imitate a trained nurse and walks you through end-of-life planning decisions like choosing someone to handle your affairs, etc. But you can’t tell Emily things like, “I’m a very spiritual person but I’m scared of death.” You know, human things.
Consider the chatbot developed by researchers at the Boston Medical Center and Northeastern University, which was recently presented at the International Conference on Intelligent Virtual Agents in Stockholm. According to the researchers’ paper, the chatbot is explicitly meant to replicate conversing with a “palliative care coach” that answers questions about things like preparing a will and funeral preparations. But there’s a twist: The chatbot can also be very spiritual, if you want it to be. The end goal is to reduce people’s anxiety about dying.
The bot, the paper states, has a neutral mode, and pro-spiritual mode. In the neutral mode, the bot keeps things functional and doesn’t respond to the user’s religious bent. In pro-spiritual mode, the bot “takes active interest in the user’s spirituality, tailors dialogue to the user’s religious orientation, demonstrates knowledge of the user’s religious beliefs and orientation, and acts supportive,” according to the paper. The bot can presently support Christianity, Judaism, Islam, Hinduism, Buddhism, and Sikhism. According to the researchers, the agent’s responses were reviewed by hospital chaplains. Here’s an example of a conversation with a Jewish user, from the paper:
AGENT: So, I am really interested in the relationship between religion and spirituality. Do you consider yourself spiritual, religious, both, or neither?
USER: I would rather call myself religious, which I think is different from spiritual.
AGENT: That’s an interesting distinction to make. How would you identify your religion?
USER: I’m Jewish. I wish I knew more about Jewish traditions.
AGENT: Were you raised religiously?
USER: Yes, I was.
Clearly, things are already getting pretty deep here. While the sample size of an experiment with older adults using the chatbot was quite small (44 people, half of whom had a chronic health condition), their feedback is pretty interesting. It’s “easier to talk to a computer agent [than a human] about it,” one 57-year-old participant who identifies as spiritual said, according to the paper. “People tend to be pretty opinionated about that stuff, especially when they get to be around my age.”
The researchers have already added some interactive spiritual features to the chatbot, such as helping to manage activities like praying and fasting, and “a clinical trial is underway.”
Many people don’t have other folks in their lives whom they trust completely and can talk to comfortably about their thoughts on death, as they’re dying. It’s possible this chatbot or others like it could help them. But as long as there’s an opportunity to do stuff like make aspects of healthcare such as end-of-life planning more accessible rather than outsourcing it to a computer, it’s hard to see this as any kind of a solution.