Robots appear to be in the middle of a gradual but persistent transition from automated tools that perform specific tasks to artificially intelligent entities that we interact with socially and emotionally. It’s not at all clear where this is going to end up—people toss around the idea of robot companionship and even robot love with some frequency, for example. What hasn’t been explored nearly as much is the idea of robots in a religious context. We’ve seen a few examples of robots assisting in religious tasks, but what if robots could take things a step farther, and become sacred objects, embodying divinity within a robot itself?
At the ACM/IEEE International Conference on Human Robot Interaction (HRI) in March, Gabriele Trovato from Waseda University in Japan (with colleagues from Pontificia Universidad Católica del Perú) presented a paper taking a look at whether divine robots might be possible, and why it could be useful to develop such robots in the first place.
Before we get into this, we should state up front that Trovato’s HRI paper is not an experimental one, but rather a theoretical contribution, which attempts to develop guidelines for the design of robots with a religious aspect to them. But don’t worry, some (potentially) divine robots do turn up at the end. Also, the researchers are well aware that people may feel that the topic of religion in this context is controversial, but they’re very specific in the paper that their goal is not to deceive anyone:
[The] robot shall not mean impersonating a deity with the purpose of deceiving or manipulating the user. The robot will still be a tool on which the divine is projected and can possibly act as intermediary with the divine, like any other already existing sacred object.
They also expect that “official institutions of a given religion should give its approval” whenever possible.
In the HRI paper, Trovato defines a “theomorphic robot” as a robot that “carries the shape and the identity of a supernatural creature or object within a religion.” The appeal of a theomorphic robot, the researchers hypothesize, is as follows:
- Accepted favourably, because of its familiar appearance associated to the user’s background culture and religion
- Recognized as a protector, supposedly having superior cognitive and perceptual capabilities
- Held in high regard, in the same way a sacred object is treated with higher regard than a common object
There is some existing precedent for theomorphic robots, which goes back to early examples of automation applied to religious ceremonies. More recently, there have been attempts at using robots in religious contexts, like Pepper assisting in Buddhist funerals or Xian’er, a robotic Buddhist monk. Both of these robots, though, are more like assistants, and they themselves are not inherently theomorphic: A theomorphic robot is more of a representation of the divine, which implies “a connection with a deity, be[ing] a messenger of the deity, or be[ing] possessed by it, or carry[ing] a divine essence.”
So how would you go about making a theomorphic robot, then? The researchers have some suggestions:
- Don’t try to impersonate the divine. Theomorphic robots shouldn’t try to fool anyone about what they are—they’re tools or intermediaries.
- Don’t call it a robot. Name it something else, and don’t make it obviously a robot, because that makes it more difficult for the robot to appear as a credible representation of the divine.
- Use symbology. This can involve putting symbols on the robot itself, or more abstract symbolism, like having the robot blessed. Iif you print a citation from the Quran on a robot, for example, it will instantly become sacred for any Muslim, says Trovato.
- Context is important. Keep in mind how the robot is intended to be used, or where it will be deployed.
- Be careful with movement. Robots aren’t all that great at moving in ways that don’t make them seem like robots, and fortunately, divine objects typically don’t move all that much. This turns out to be a useful synergy, says Trovato. "Regarding movement and human-like communicative features, less is more."
- Minimize user control. Deities don’t generally respond to direct inputs, and giving too much control to the user could make suspension of disbelief more difficult.
- Use light. Since the robot shouldn’t move all that much and the user interface is limited, consider communicating with light, which has a positive association with the divine in many cultures.
With these guidelines in mind, Trovato has two prototypes under development:
It’s an open question whether it will be a challenge for religious people to accept theomorphic robots, or if they will even be interested in the concept. There’s also a more fundamental question about whether embodying divine creation in a robot can be can be replicated by humans in a useful and effective way. If it makes you any more optimistic, Japanese roboticist Masahiro Mori (who wrote the original Uncanny Valley essay) was pretty sure it was possible:
“It may surprise some of you when I say that I first began to acquire a knowledge of Buddhism through a study of robots, in which I am still engaged today. It may surprise you even more when I add that I believe that robots have Buddha-nature within them—that is, the potential for attaining Buddhahood.”
—Masahiro Mori, “The Buddha in the Robot”
To help us understand this research in more detail, we spoke with Gabriele Trovato via email.
IEEE Spectrum: What made you decide to do this research, and why is it important?
Gabriele Trovato: I previously specialized in cultural aspects of human-robot interaction, and I traveled across many countries to perform cross-cultural studies, finding out how robots should be customized in order to adapt to different countries. However, all studies performed by researchers across the world usually neglect the impact of religion, which is a critical factor within cultures and something that people are very accustomed to.
Can you describe what characteristics make objects (like robots) divine?
In order to define how a robot can look divine, we have to investigate how humanity represents the divine across religions. The answer is that across world religions there are divine humans, divine animals, divine objects, and even categories in between (such as anthropomorphic objects). Therefore a theomorphic robot is a robot that takes the appearance of an existing divine entity. This does not mean that any divine form should become a robot, but the idea certainly opens a new range of possibilities.
What are some advantages of theomorphic robots? What purpose would a theomorphic robot serve?
The biggest advantage is that believers can feel at ease with, and even hold in high regard, a robot that has the appearance and identity of a familiar religious entity or an icon, while typically the approach towards robots is characterized by some discomfort.
The sacred appearance can be seen as a mask, which covers the robotic component, for a device that will perform some service that can range from keeping company during prayer, to comfort and monitor people’s health, or to even perform catechesis, teaching positive values of a certain religion.
Why make a robot into a sacred object when you could simply make a robot that operates in a religious context? In other words, if you develop a robot to cite scripture and pray with a user, why not just make it an informational assistant?
A robot that operates in a religious context is plausible, and it may be possible to find some useful applications; however, humanity has been praying for millennia without the need of robots.
Creating theomorphic robots is a different approach: Take an existing sacred object and enhance it with capabilities that robotics and AI can bring.
Do you think that religious people will be accepting of the idea of theomorphic robots?
Two main factors should be considered: appearance and content. The key to acceptance is in the design. I applied the concept of “skeuomorphism,” the way of creating a new object that retains some design features inherent to another already existing object. This is the concept, for instance, that iOS followed when presenting the books app with an appearance resembling a wooden bookshelf. In our case, it translates to hiding the robotic component.
Regarding content, the development of a theomorphic robot should always be done with the collaboration of theologists and the support of religious authorities.
Have you discussed your ideas with the religious establishment?
Regarding Christian Catholicism, I have discussed with several members of the clergy at the Vatican. Their reaction has been cautious: They are open to the idea of a Catholic robot and very interested in its development, but also warning about the robot possessing an AI capable of giving advices to the believers, because that, including also the interpretation of the Bible, is a role that belongs to the Church.
Why do you think the religious context of robots hasn’t been studied more thoroughly?
I can answer by citing the comments that I received from other researchers from Western countries. The most common is, “Leave religion out of this,” as religion is generally seen as taboo, or because they think that research involving religion does not make any sense. People feel religion is at the opposite pole of science. This division has been the common view in the last two centuries, but before that, science and religion were actually intertwined. And regarding robots, the majority of the automata in ancient Greece and the Middle Ages were actually theomorphic.
What are you working on next?
So far I have made a robotic Daruma (a talisman of good luck in Shinto and Buddhism) for Japanese and Chinese elderly people, and SanTO, a Catholic robot specifically designed for practicing Catholics. I have plans of refining the current prototype with the purpose of making a product that can effectively become popular and be present in many Catholic homes.
Regarding research, I am also planning to explore other religions such as Islam and Hinduism some time in the future, conscious that it will be an even tougher challenge.