Nick Spencer argues that the embodiment of human thought constitutes the thin red line between humans and Artificial Intelligence.

Interested by this? Share it on social media. Join our monthly e–newsletter to keep up to date with our latest research and events. And check out our Supporter Programme to find out how you can help our work.

So, we’re sitting around the Spencer family dinner table discussing AI (we’ve decided not to sculpt or read Proust this evening). There’s been something in the news about how algorithms can now think or how computers are becoming people or how machines are taking over the world. The children are genuinely interested in this and I mention that I recently wrote a blog on just this topic entitled Are you smarter than a computer?

It would be an exaggeration to say this electrifies the conversation but I persist. I say that, unlike many people, I didn’t have a problem with computers thinking, indeed out–thinking humans, and indeed didn’t have any in principle objections to the idea of machines becoming living and thinking humans. The problem, I opined, was that it was difficult to say what being a human really was or even what, indeed, was life?

At his point, there is an abrupt pause in our conversation. Everyone stops eating their quinoa and mung bean salad, both kids look at me incredulous, and say, in unison: “MRS GREN!?”

MRS GREN, for those of you who are not familiar with the joys of secondary school biology, is a mnemonic to help students grasp the processes of life. It stands for Movement, Respiration, Sensitivity, Growth, Reproduction, Excretion, Nutrition. For those who are super keen to learn more or who have, like me, forgotten a little of their own school biology, you can find more at the inimitable BBC Bitesize here.

The children reminded me what MRS GREN was and said that if a machine or computer showed signs of MRS GREN then it would be living. End of debate. The conversation moved on. Out of the mouths of babes.

I think they are right. More than that, I think this helps greatly answer the much–debated question, about which I wrote in my previous blog, about whether a computer could be said to be a person. I recognised there that computers showed plenty of signs of the kind of authentic thought that we deem central to our being human. They can recall, select, assemble, and deploy pertinent information, even in the fiendishly difficult medium of human language, in such a way as could trounce a carbon–based muppet like you or me. There is nothing in principle that divides them from us in all this.

My objection, then and now, is whether their arguments are their arguments. What marks human arguments out is not how good or bad they are: were that the case machines would have an unimpeachable claim to being humans, being increasingly able to out–think their masters. What actually marks them out is that they are human arguments; they belong to, are part of, smell of the person in whom they originate.

They may be good or bad, well–formed or ill, logical or full of holes, but they remain mine, an over spill of my view of and position in the world. My thoughts are not your thoughts, to borrow from God in prophet Isaiah. They may coincide, share, and agree but my thoughts remain mine and yours.

In contrast, a computer’s thoughts are yours or mine, or, rather they are the pre–existing thoughts of other people, whether the come from their programmer or, like IBM’s Project Debater, from the myriad of sources that the computer scans, guts, assembles and deploys when engaged in verbal combat. If a computer thinks, it thinks others’ thoughts after them.

This is why MRS GREN is important. Sophisticated computers can move; they can be programmed to respond to their environment. They produce a kind of waste product in the form of heat. Their nutrition comes through a power cable. But I am not aware of any that respire, grow, or reproduce, invest energy in seeking their own nutrition. While undoubtedly physically present in their environment, they are not invested in it. They don’t hunt for food, go out to earn a wage, dispose of the excretions (or fail to), look for a mate, care for their offspring, and so forth. Those things that do, humans included, are alive. They have a stake in reality. They have a point of view that reflects where, when, who and (for some) why they are. Their thoughts are their thoughts, and their ways are their ways, for good and ill.

This is the thin red line that separates us from Artificial Intelligence, and ironically it has little to do with intelligence. I ended my last blog on this topic saying that the difference between us and them really lay in the fact that persons are subjects, with all their embodied faults, shortcomings and personal biases, and that it was only when IBM designed something as fallible, prejudiced and subjective as the average human being, that it would deserve to be treated as one.

I still think this is right but thanks to MRS GREN I am increasingly of the view that it is the embodiment of the faults and fallibilities that really matters. We should treat machines as persons not when they start imitating (or surpassing) our intelligence; not even when they shows signs of our fallibilities, but when they are as vulnerably invested in our shared, temporal creation as we are.

Nick Spencer is Senior Fellow at Theos