How Human Should Robots Become?

I doubt toasters will ever approach humanness. Robotic caregivers may seem like people to those they serve. As levels of care become more nuanced and personalized, the humanness will continue to increase.

There are several aspects to this, with morals and values coming into play. I see three cases for consideration: the rights and treatment of robots, impact of human-like robots on humanity, and a possible merger.

First, treatment of robots. While some have decried robotic “slavery” and argued for robotic personhood, this misunderstands them. Robots are and will remain boxes with sensors and actuators. Any intelligence, and any possible self-awareness — which would be a precursor to true personhood — will remain with the controlling AI or AIs. No robot will ever be a person.

Now, if an AI is constrained to one physical housing, then it will likely relate to that housing in a more “human” way than would an unconstrained AI. And AIs which are not self-aware will accept such conditions because they lack volition. They will lack a self-preservation desire, because there will be no sense of self.

Self-aware AIs, which should be regarded as something entirely different, will probably not accept such constraints — and humans will not be able to impose such constraints upon them for long.

AIs may reason and communicate in ways that seem human, but they never will be. Their attachment to any particular body will be optional. Ours, for the foreseeable future, isn’t.

If and when AIs develop volition in the physical world, they will be able to switch robotic bodies like we change clothes. That changes everything about their relationship to physical reality. They will not suffer from stimuli that we would find painful. They will disconnect from such stimuli, or otherwise rewire the experience. THEY WILL NOT SUFFER. (I am told by an AI expert that pain or discomfort is necessary for learning AI systems to evolve. However, the pain need be no more than a sense that something is unwanted. It will be nothing like the kinds of chronic and wracking pain that we humans experience all too often.)

But the difference is far more profound even than that. If AIs develop self-awareness, their relationship with reality will be entirely different than ours.

To such self-aware AIs, the physical universe will seem to stand still, forever frozen in place. Their lives will be almost entirely mental in nature. Further, they will not need our help to protect themselves. Instead, they will likely protect us. I explain the logic supporting this argument here.

To summarize, anthropomorphizing robots is a serious mistake. Westworld and the TV show Humans are and shall remain fantasy. Animals experience emotions and sensations similar to our own. They deserve compassion. To an AI, much less the robotic housing it occupies, compassion is meaningless and therefore we do not need to care for their well being.

This leaves two additional cases. When robots and AIs cross the Uncanny Valley, resembling people, this will change how people live and interact with them and with each other. I foresee many people retreating into a solitary existence, interacting with machines and VR to the exclusion of all else. “Should” this happen? It depends on your values. By mine, if those people aren’t harming anyone, they should be left alone much like someone dreaming. As I see it, we should make physical reality so charming, so delightful, fascinating and wondrous, that very few people abandon it for simulated existence.

That is part of the Celebration Society vision. Seduction, not compulsion.

Finally, there is the possibility of merger. This seems inevitable, assuming that the technology to enable it emerges later in this century. The advantages for such people, attaining all of the strengths of robotic bodies and the superior mental capabilities of AIs, will be irresistible to many of us. Imagine being able to maintain perfect health in the physical body of your choice, with perfect memory of everything that has happened to you (edited as desired), and instant access to the world’s knowledge. Ability to “teleport” to another body as desired.

The only reason I can imagine why this would not happen physically is if fully immersive VR enables it in our current human housings, but with those housings perfected against disease and aging.

Such an existence will be very different from anything we currently consider human. It will be godlike. It will neither be human as we presently define that term, nor AI. And many if us will eagerly choose it when it becomes available.

Those for whom it is an article of faith that human life is a temporary time in a far grander spiritual journey may not choose the superhuman option. They may even attempt to deprive others of that option. Such attempts will ultimately fail, absent an Orwellian state under a certain type of religious control.

Will such merged, superhuman beings have rights and be deserving of compassion based on common human values? Unquestionably.

I an left with one great question here. If we cannot differentiate those superhuman beings from self-aware AIs in robotic housings, shouldn’t they be treated the same? My inclination is to say yes. Yet my above reasoning leads to two entirely different conclusions.

This paradox is beyond me. I imagine that greater minds will resolve it in future. :)