Yes, machines are starting to think for themselves, but do they have a soul?
By Catherine Pepinster
Scientists call it the second machine age. The Industrial Revolution ushered in the first machine age, in which machines replicated human physical labor.
Now, the Second Machine Age will see human mental abilities replicated and even enhanced, thanks to advances in artificial intelligence (AI), a sophisticated form of computing that replicates human cognitive functions.
AI is now developing at such an increasing rate – and being used to solve specific problems – that the second machine age is almost upon us. Machines can perform tasks until recently reserved for humans, such as detecting disease, driving a car, forecasting a restaurant’s food needs or optimizing the logistics of a global retailer. The AI does all of this by identifying patterns in the data. It is a technology inspired by the human brain.
So if AI is developing so rapidly – cars that can now park, robots that provide care for people with dementia, artificial limbs that can run and dance, and drones used in war – how do these machines look like us? Could robots become so advanced and replicate people that they might even one day have souls?
That was the question posed during a briefing, Spiritual Silicon: Could Robots One Day Have Souls?, co-hosted by the Religion Media Center and Theos, the Christian think tank. This was the first of three sessions on science and religion, drawn from the Theos research project, Science and Religion: Reframing the Conversation.
The research found that the vast majority of people (75%) weren’t convinced that robots looked like us: they didn’t believe they had a soul. And yet, despite this convincing rejection, a small percentage (5%) thought they could do it, given their intelligence abilities.
What is it about robots, as smart as they seem, that makes the majority perceive them as soulless? In other words, what makes man human? The members of the information panel focused not so much on our abilities as on the exact opposite.
Laura Janner-Klausner, rabbi of Bromley Reform Synagogue, suggested that people have a deep vulnerability, while Nick Spencer, principal researcher at Theos, ventured that our physical body – the thing that can be so vulnerable – is the main vital difference.
Anthropologist Kathleen Richardson, professor of robot and AI ethics and culture at De Montfort University in Leicester, said: “Human beings do not come into the world fully formed. So humans need relationships.
These arguments have helped us understand how humans are different from robots, but what about the soul? Theos kept his soul questions in his investigation non-specific, instead focusing on the clarity of the question.
AI specialist Professor Neil Lawrence said: “It’s [about] a fundamentally different intelligence – machines are not embodied. However, he warned that dealing with issues surrounding the soul is more difficult because people keep moving between different words such as sentience, sapience and conscience. But he noted that many of them derive from a Latin word meaning to breathe.
This is certainly related to some attempts to define the soul as the spark of life in each person, or their deepest being. Could a machine ever be considered to have such a spark?
Rabbi Janner-Klausner explained how we even deal with objects that impact our humanity. If a robot cared for a parent with dementia, say, should we just pull the plug when it’s all over? “Should we bring a ritual around her?” she asked. “If your mother is taken care of who has Alzheimer’s disease, and you put her in, you bring your violence to her, it encourages your violence. It’s not about the robot. It’s about you.”
But others worried that the problem with AI and robots is that we make too many assumptions that they look more like us than they really do. “We have to be careful of anthropomorphism,” Prof Lawrence said, and Prof Richardson said: “We’ve locked robots into anthropomorphic language, but no machine is like us… We’ve immersed ourselves in a disturbed fantasy about machines.”
Dr Spencer argued that our history showed humans make mistakes about what constitutes a person – he cited racist attitudes towards black people – so it was worth questioning what we were looking at. He reminded the briefing that intelligence and processing power were often cited as what characterized a human being. But many humans – like those born with profound disabilities or those who have suffered debilitating illness or terrible accidents – lacked these abilities. Weren’t they human? If they all had a soul, it helped to define them still as human.
Mark Vernon, who is part of the Understanding Spiritual Intelligence research project, run by the International Society for Science and Religion, said that instead of having souls, robots offer something to our own souls: akin to relationships people’s personal belongings with religious objects, such as icons or Bibles? These objects are inanimate but resonate with our soul, which is why we treat them with respect.
He added: “No computer has ever felt the slightest sense of amazement at its activities, as we regularly do.”
Eve Poole, Acting Chief Executive of the Royal Society of Edinburgh, quickly fired back: “We don’t know that!” Dr Poole, more enthusiastic than others when briefing for the potential of AI and robots, added: “We’ve seen self-awareness in machines that are quite simple.”
David Midgley, a Buddhist and founder of the fledgling Jamyang Buddhist Center in Leeds, wasn’t so convinced: “We don’t know anything about the subjective experiences of machines, or whether they have any,” he said. “Therefore, we have no reason to believe that they do.”
The Theos research team, in collaboration with the Faraday Institute for Science and Religion and funded by the Templeton Religion Trust, interviewed 100 experts and commissioned a survey of 5,000 adults, and found clear differences in opinion according to age and religious beliefs.
The study showed young people are more receptive to the idea of extending human rights to robots, with 27% of those under 30 believing it will happen one day.
The study also found that those with more fundamental religious beliefs, Christian or Muslim, were more likely to believe that one day robots would have souls – presumably as they progressed in their power and abilities, making them look more like thinking humans.
He also found that:
- Those who pray frequently are slightly more likely to disagree with robot souls (80%) than those who pray occasionally (76%) and those who never pray (74%).
- The groups most likely to be against the notion of robot souls are those who are clearly against the idea of human spirituality or immortality – followed closely by those clearly for human spirituality or immortality.