Hypothetical. We have superintelligent, bipedal nurse robots in the 2050s. What attributes would they have

So assume by the 2040s or 2050s, we have bipedal nurse robots which have replaced humans. What attributes would be good for a nurse, or any other health care worker, to have?

Obviously it would be super-intelligent, so it would have a perfect understanding of every scientific paper and book ever written on subjects like human biology, biomedical engineering, medicine, pharmacology, etc. It would understand the hundreds of millions of books and scientific papers ever written on these subjects, and be able to understand the errors in them too. It would also have a perfect understanding of the patients entire medical history, a full understanding of all 3 billion base pairs of their DNA.

It would obviously work 24/7, and probably have an exchangeable battery pack that gets swapped out every day or two so it doesn’t need to go down to recharge, you just replace the battery pack (the robots would probably have some internal reserve battery so they could replace their own battery packs without shutting down).

What I’m wondering is what physical attributes would a robotic health care worker have?

For example, why not give it six arms? It seems like when a patient is experiencing something intense like surgery or trauma you need multiple medical professionals to each do something different. A robot with 6 arms could in theory do CPR, insert an IV, put pressure on a wound, and prep equipment all at the same time while now that may take 3 or 4 medical professionals to do.

Eyes which have a far better capacity to detect shade would be nice, perhaps to pick up on subtle shades of color in people’s blood, skin color, etc. So superhuman senses like vision, sound, sense of touch, smell, etc to pick up on things that human beings cannot pick up on. Perhaps a robot nurse could hear your breathing change from 100 feet away and know a medical emergency was about to happen.

Would hands with 4 fingers and a thumb be ideal? I’m assuming there are physical interfaces to interact with the patient and interact with equipment better than primate hands that would work well in the medical profession, but I’m not sure what would be superior.

An in-built port to do lab diagnostics. Maybe something to read various POC tests like antigen tests, antibody tests, etc. I don’t know what all lab tests can be miniaturized so that a robot could carry them on them though so they could test themselves rather than taking samples to the lab for testing.

I hope I’m not the only one who immediately thought of this sketch…

We will have robots for that too I assure you.

They will have the attributes you specify when you order one.

Are these nurse robots, or doctor robots? Because most of the traits you describe would be good for doctors, but irrelevant or actively detrimental for nurses. Nursing is all about the human interaction.

For a robot that would be good at nursing, go with Baymax.

Do you think the human interaction will really matter for the majority of nursing though? Because AI is going to have a virtually perfect understanding of psychology and sociology as well. They will understand endless millions of books and scientific papers written about how human cognition and emotions work, as well as be able to pick up on subtle ‘tells’ about a person’s inner psychology that even their closest friends and family don’t know about.

I could see someone in hospice or afraid of death wanting human interaction, but a lot of nursing isn’t about dealing with people literally on death’s door.

Also there is the economics. When a bipedal robot that works 168 hours a week costs $25,000 but a human nurse that works 36 hours a week costs $90,000, hospitals will choose the robot. Also (in theory) some of those cost savings will be passed onto consumers which means consumers and insurance companies will prefer the robot medical professionals too.

Also a lot of patients will prefer the higher competence levels of the robot. There is a lot of gaslighting in medicine (trying to blame everything on obesity or anxiety), and humans can’t keep up with the information overload.

Also I meant all medical professionals. I guess I just picked nurses because I meant a bedside medical professional the patient interacts with the most.

That’s like asking whether handwriting really matters for a scribe, or whether athletic ability matters for a pro football player. Human interaction isn’t a component of the job of a nurse; it is the job.

Encyclopedic knowledge of medicine is great for a doctor. It doesn’t matter for a nurse, because that’s not a nurse’s job.

I’d imagine that the patients would be wired up with sensors and have their IV instrumented such that when something got out of whack, the robo-nurse could just automatically correct for it without having to even move.

In fact, I’d think the robonurse would be mostly a robohousekeeper that would clean bedpans and change sheets, and move patients. There wouldn’t be so much need for an ambulatory robonurse if the patient was adequately instrumented.

I don’t think that’s true. When my father was in intensive care, his primary care giver was a nurse, named Tom. My father was unconscious, and at death’s door. There was little or nothing in the way of “human interaction”. Tom monitored my father’s vital signs, and was constantly fiddling with IVs and dials. He replaced components of machines. He was always moving. At some point my dad improved enough that they moved him to regular intensive care, and he got a new nurse, who did interact with him some as a human. But she mostly tended to his extensive physical needs, like making sure the breathing machine was working at the right speed.

You’d still need nurses and CNAs to change sheets, assist with using the bathroom, doing CPR, taking blood, doing physical exams, provide food, etc.

Have you met many nurses? They have advanced knowledge of medicine. Sometimes better than the physician.

The term “nurse” covers a lot of ground. My recent GF is an RN in a trauma OR. Her tasks there include arranging the barely alive usually unconscious body on the table for best access, cleaning it up, covering it up except for the operative sites, dealing with the plethora of tools the surgical team needs, being the data logger and photographer for everything they do, assisting each of the several docs in their hands-on work, handling the admin of transferring responsibility in and out of the OR team, etc., handling the recovering patient off to the recovery team, or sorting out sending them down to the morgue.

A typical operation, to the degree there is “typical”, has all those tasks divided among 2-4 nurses. But by and large it’s technical deliberately distant unfeeling work dealing with a mangled semi-corpse that you’re trying to pull back from the brink.


A “nurse” is also somebody who cares for the infirm elderly; bathing, watching, talking with, checking up on, helping them walk or eat or, etc. Which is an intensely human endeavor the polar opposite in my GF’s job in terms of the skills and aptitudes required.

As to robots, for this latter job it’s not enough for the robot to be empathetic (or a skilled mimic of empathetic). It must be seen and felt by the patient to be empathetic. That’s gonna be a tall order; we here are the target market for 2050 elderly infirm. How do you feel, not think, about a robotic “nurse” feeding you your oatmeal?

Conversely, the knowledge, technical skills, extra appendages, and calm courage under fire the OP was mostly going on about are just the ticket in the trauma OR.

I disagree a little on this. I think it’s pretty natural for humans to anthroporphize. I find myself praising, or being annoyed by, or thanking Chat GPT often despite knowing it can’t love me back :smiley:

That’s not to say an AI can entirely take the role of a councellor or friend; people who are very upset may still need that human interaction complete with a warm hug. But IME a lot of people just want someone to listen to them and give them attention, and IMO chatbots are already better at this than many humans.

Heck I think there’s already a great use case for a machine that just does:

  1. able to carry an elderly person from one place to another, and pick them up from the ground
  2. with chatty LLM

and the reason this doesn’t already exist (or isn’t widespread yet) is just it’s a PITA to license new medical tech. And stairs.

The only reason to make NurseBot bipedal is so it looks human and presumably puts patients at ease. Is you’re going to give it six arms and interchangeable wrist attachments, you’d be better off having the bot on wheels, tracks or floating around like a Fallout-style Mr. Handy than a humanoid spider-monster.

Actually, a spider-monster appearance could be a positive drawback, particularly where pediatric nursing is concerned.

They will be owned and programmed by the same profit-guided corporations that currently own our health systems, but they won’t have any of the ethics/morals/sympathy etc. that human nurses may have.

Yes, “bedside manner” was an immediate concern for me when I saw this thread. I do think “polite concern” speech and responses can be programmed into the devices, but I still believe humans cannot be entirely cut out of the nursing process.

The second they can be, they will be.

I think it would go over well in psych nursing

Not quite.

In robotics research it’s become real obvious that if you’re interacting with an environment designed for humans, the more human-shaped the robot is the better it fits in the space it’s sharing with the humans.

Witness all the accommodations we choose to build into our buildings for the wheelchair users. And even then, most of the built environment is still very much inconvenient for them since it wasn’t designed with their shape in mind. It’s easier to build robots with legs able to climb stairs or stools than it is to build or re-build the facilities they work in to accommodate their tracked bases or whatever.

Right now there are legit mechanical engineering and power consumption reasons that make a wheeled or tracked base a LOT easier to build than a bipedal design. So far.

An interesting tradeoff to be sure.