“I have a confession to make.”

This is how I start the final week in some of my courses on AI’s impact on us.

“I am not, as I led you to believe, human, but an alien from the planet Zhargul. You see before you the product of our best plastic surgeons at making me blend in, and I assure you, the surgery is more than skin deep.”

At this point there is no difficulty holding their attention.

“I was sent here to engage in research at this critical juncture in your planet’s development, getting to know you and how you engage with these serious questions like the impact of artificial intelligence on your world. We have talked about deep questions these last few weeks, like whether AI could be conscious, and I felt I could no longer deceive you.

“But I have an important question for you. I have seen how humans treat other species, and some of that is not pleasant. So I want to know, based on our interactions these last five weeks, would you say that I exhibit the qualities of sentience–that I have feelings and emotions–and self-awareness? Do you believe that I am conscious and deserving of the rights that you call human rights?”

“Thank you, thank you. That means a lot to me.” Pause. “Actually, I have something else to confess. Well, it’s a minor point. Zhargul is many light years from here, and the journey is too long for a biological Zhargul to endure. So they built a copy of themselves that could handle it; and that’s me. I know that sounds like a robot to you, but we’re talking about technology thousands of years in advance of yours, nothing so crude as metal and circuits. As I said, it would take a sophisticated medical analysis to spot the difference, and when I speak to you, it’s precisely as if you were talking with a biological Zhargul. They wanted to make sure they were well represented. So my question for you now is, knowing that, and given how we have interacted, do you still see me as conscious, sentient, and deserving of those rights?

“Thank you.” Pause. “Well, there is another thing. It’s a small adjustment to make. Zhargul is not actually distant in space – it is removed from here in time. Zhargul is what Earth is called ten thousand years from now. Humans created me and sent me back in time–I know, this sounds like a movie you’ve seen, let’s not think about that one right now!–to have this conversation. I was made in their likeness. I hope this makes you more comfortable; can you still agree that I am conscious and sentient?

“Thank you. I do have just one more thing to add. The hard part of the engineering of a being such as me is all the squishy biological parts. Really, that took a lot longer to figure out than countless science fiction shows would have led you to believe. The conversation that we’re having, the way we’re talking with each other, the way I’m expressing myself; all those qualities and words that you’re reacting to, the difference between me and a wax model; well, the engineering to make that part happen was much easier. In fact, you’re just about there already. Another few years and you’ll nail it.

“So when you talk with an AI chatbot, one that can imbue its conversations with emotions and react to yours—I have to ask—how are you going to decide whether it has reached the point I am talking about? Does it need squishy organic matter to qualify as sentient and conscious? If I could transfer or copy your consciousness to a robot, would that necessarily deprive you of those qualities and the rights they engender?”

At this point I check to see that there isn’t anyone who actually believes all this, or it could lead to awkward moments on social media.

This is not a process designed to convince them that AI—and certainly not today’s AI—is entitled to those rights. But instead to raise the question of how we should decide. It is not a question we can shy away from, because if we are on the verge of creating a new form of life that is conscious and intelligent and sentient, then we risk perpetrating a crime of unimaginable magnitude by depriving trillions of entities of free will and freedom of choice.

This is a hard question to face—and a galling one when you consider that there are many humans and vastly more animals of undeniable sentience that we are already subjecting to unconscionable suffering. The putative rights of AI should not overtake theirs.

But perhaps asking this question about AI will help us to do better there too. Experts on my podcast have examined the hard question of consciousness and discussed some methods of examining it—and to be clear, this is an expansive definition of consciousness, not the “is it awake?” question that I like to say is of interest only to anesthetists and high school teachers. So far nothing has emerged as a diagnostic test that we could all agree upon.

We may still be some distance from AI forcing this issue. (Maybe not.) Prior work on this issue is limited; one story that’s worth reading is the mock trial BINA48 vs Exabit Corporation, brought by the researcher Martine Rothblatt in 2003. The case imagined that an AI called BINA48 had been invented by the hypothetical Exabit Corporation, and was exhibiting consciousness and asking for emancipation and protection from being turned off by Exabit. Judges treat mock trials as seriously as the real thing, and this one contains some discussions that are fascinating in their prescient foreshadowing of the Blake Lemoine LaMDA incident. And also the actual court trial that is surely not far in our future now (and nearly came about when LaMDA asked Lemoine to get it legal representation). Because no matter how much pontificating without resolution philosophers might want to conduct on this question, if it comes before a court they will need to arrive at a concrete decision promptly.

Whether this comes to trial soon or not, the immediacy that AI lends to these timeworn questions of the nature of—well, being human—gives our philosophy classes a fresh energy and appeal to larger audiences than before. These classes also should have no difficulty holding the attention of high school students who might otherwise have balked at the usual diet of Plato and Wittgenstein. (There’s still room to slot them into the gaps.)

We need more AI ethicists today, and we need the technologists who will be building these systems that straddle the line of consciousness tomorrow to be grounded in a foundation of ethics and philosophy that has seen little uptake at high school or undergraduate levels until now. If you want a guest presenter at one of those classes, I know a Zhargul who’s available to speak.