In today’s column, I examine an intriguing twist to the usual discussions about the use of AI for mental health advisement.
First, people are increasingly making use of their PCs and smartphones to access generative AI and ask questions about their mental health, essentially using AI as a kind of therapist. That’s an abundant fact and widely reported. Second, the upcoming twist will be that humanoid robots, physically resembling humans, will be improved sufficiently so that they will adroitly function in our homes, including being closely paired with the latest in generative AI.
This means that people will be able to readily access generative AI-driven mental health advice from these human-looking walking-talking robots. How will that spur people to seek mental health guidance via AI? Some pundits insist that people will be even more likely to confer with a humanoid robot, while others say that people will be creeped out, thus decreasing interest in conferring with AI about their mental health.
Let’s talk about it.
This analysis of AI breakthroughs is part of my ongoing Forbes column coverage on the latest in AI, including identifying and explaining various impactful AI complexities (see the link here).
AI And Mental Health Therapy
As a quick background, I’ve been extensively covering and analyzing a myriad of facets regarding the advent of modern-era AI that produces mental health advice and performs AI-driven therapy. This rising use of AI has principally been spurred by the evolving advances and widespread adoption of generative AI. For a quick summary of some of my posted columns on this evolving topic, see the link here, which briefly recaps about forty of the over one hundred column postings that I’ve made on the subject.
There is little doubt that this is a rapidly developing field and that there are tremendous upsides to be had, but at the same time, regrettably, hidden risks and outright gotchas come into these endeavors too. I frequently speak up about these pressing matters, including in an appearance last year on an episode of CBS’s 60 Minutes, see the link here.
If you are new to the topic of AI for mental health, you might want to consider reading my recent analysis of the field, which also recounts a highly innovative initiative at the Stanford University Department of Psychiatry and Behavioral Sciences called AI4MH; see the link here.
Social Robots Versus Humanoid Robots
Let’s first explore the nature of so-called social robots and then consider the upcoming advent of more advanced robots that are referred to as humanoid robots.
You might be familiar with social robots, which are today’s commonly used robots for home use.
Though they don’t especially look like humans overall, they typically have a vague resemblance by showcasing some human-like features. For example, there are ball-shaped robotic heads that look somewhat like a human head but lack the fluency and realistic appearance of humans. The robotic head lacks a human-looking body. It has crude-looking mechanical eyes and a rudimentary mouth, all of which move in a jerky manner.
There is little doubt that you would instantly consider such robots as being simplistic, cute or quaint, and pretty much viscerally non-threatening or otherwise unimposing.
Meanwhile, we are heading at a fast clip toward much more advanced humanoid robots.
Humanoid Robots Are Coming
A humanoid robot is a type of robot that is purposely built to resemble a human in complete form and function. There is a robotic head, arms, hands, fingers, legs, body, etc. You’ve undoubtedly seen those types of robots in the many online video recordings showing them walking, jumping, grasping at objects, and so on.
A tremendous amount of active research and development is taking place to perfect humanoid robots. They look rather comical right now. You watch those videos and laugh when the robot trips over a small stick lying on the ground, something that a human would seldom trip over. We scoff when a tested humanoid robot tries to grasp a coffee cup and inadvertently spills most of the liquid java.
It all seems humorous and a silly pursuit.
Keep in mind that we are all observing the development process while it is still taking place. At some point, those guffaws of the humanoid robots will lessen. Humanoid robots will be as smooth and graceful as humans. This will continue to be honed. Eventually, humanoid robots will be less prone to physical errors than humans are. In a sense, the physicality of a humanoid robot will be on par with humans, if not better, due to its smoothly refined mechanical and electronic properties.
Do not discount the coming era of physically capable humanoid robots. This is going to have a massive impact on society. For more details on AI and humanoid robots, along with research known as physical AI, see my in-depth discussion at the link here and the link here.
Pairing AI With Humanoid Robots
AI has a lot to do with humanoid robots, including these two major facets:
- (1) Operational AI. There is a mechanical-oriented AI that aids the fundamental operating aspects of a humanoid robot in its core functions of walking, running, handling objects, and so on.
- (2) Language AI. There is generative AI that enables humanoid robots to be seemingly fluent in natural language and interact conversationally with humans. That’s the talking side of humanoid robots. Note that generative AI can also direct onboard operational AI, too.
I will briefly unpack those two facets.
Humanoid robots lean into operational AI to aid in giving the robot a bunch of very fundamental or barebones capabilities. There is AI that does vision processing and analyzes the images being captured by the embedded cameras that are considered the eyes of the robot. The ears of the robot are recording sounds that the robot hears and feeds the audio into a specialized AI that analyzes it.
All in all, a whole lot of AI is working inside the robot to enable the robot to maneuver in the real world.
Beyond those core functional aspects, generative AI is being added as an integral element of humanoid robots. Here’s how that works. A customized generative AI or large language model (LLM) is placed inside the processors and data memory contained within the frame of the humanoid robot. This will provide a self-contained generative AI or LLM that doesn’t need any Internet access.
We might also include Internet access so that the generative AI can reach the Internet to get additional information, but we don’t want the robot to be dependent on network accessibility. The idea is that the humanoid robot needs to have immediate access to the LLM and not be delayed if network connections are spotty.
Right now, fitting a full-sized LLM into the robotic processors might be prohibitive, so an SLM (small language model) might be used instead. See my coverage on the emergence of SLMs at the link here.
Humanoid Robot As Therapist
What will a humanoid robot that contains generative AI be able to do?
The robot will interact with humans in the same manner that you use generative AI on your PC or smartphone. You normally and readily chat with generative AI, and it responds accordingly. Easy-peasy. The same applies to generative AI inside a humanoid robot. Think of this akin to using Siri or Alexa, namely, you speak to the AI and the AI speaks to you.
A tremendous difference is that the generative AI can engage the rest of the humanoid robot to respond.
You might tell the humanoid robot that you are thirsty. Generative AI analyzes your comment and tells you that it will get you a drink of water. At that juncture, the generative AI conveys to the operational AI that the robot needs to walk over to the sink, grab an empty cup, fill it with water, and walk it back to where the person is.
All kinds of tasks will be performed. Routine tasks can be undertaken, such as doing laundry by having the humanoid robot take your clothes to your washing machine and dryer, and other tasks such as cooking food in your kitchen. The sky is the limit. Whatever task a human would conventionally carry out in a home, the odds are that a humanoid robot can do something similar.
The kicker is that we can also expect that humanoid robots will be your in-home therapist. Just as you might seek mental health advice from generative AI on your smartphone, you can do the same with the humanoid robot.
Boom, drop the mic.
Myriads Of Scenarios
Some envision this therapeutic process entailing you in your cozy living room, residing on your favorite couch, and the humanoid robot sitting in a nearby comfy chair. There you are, acting as though you are in a therapy session like the days of Freudian analysis. You pour out your heart to the humanoid robot. The generative AI inside the robot responds by speaking to you via the robot.
Tell me more, the humanoid robot says.
That’s not necessarily how things will go. The effort will be much more seamless and integrative. Though you can certainly opt to do a kind of staged therapy activity, the chances are that the discussions will be more fluid and shall we say naturally or spontaneously arise.
For example, you might be walking around your home, and the humanoid robot is doing so with you. You decide to point to the laundry basket and tell the robot that it should do the laundry later that day. At that moment, you make a quip that your roommate doesn’t seem to be willing to put their dirty clothes into the basket. This mindless act by your roommate really irritates you, and you are quite angry with your roommate. Exceedingly angry.
The humanoid robot then politely asks you if you’d like to chat about your brewing anger. You say yes, you would. You and the robot continue to walk throughout the house, doing various chores, and simultaneously discussing your anger. Turns out your anger is more than the mere act of your roommate leaving clothing around the house. It is much more deeply rooted, and the humanoid robot helps you cope with the unbridled anger.
This discussion happens seamlessly and with a natural kind of rhythm.
Welcomed Or Repulsed
Wow, that’s great, some might proclaim. You will have an AI-based therapist at the ready, conferring with you whenever you want. No need to do any kind of login or take overt actions to dive into a mental health discussion. It will happen smoothly.
Sometimes you will initiate such a chat. Other times, the humanoid robot will do so. The laundry situation is a prime example of this. The person mentioned their anger but didn’t directly tell the AI to engage in a devout therapeutic mode. Instead, the generative AI analyzed what the person said and offered that perhaps a therapeutic conversation might be helpful. The AI was proactive in that circumstance.
The reaction by some is that they would never interact with a humanoid robot in this way. It seems creepy. Maybe even a bit scary. You see, opting to reveal your innermost feelings is something sacred, and doing so with a robot is zany.
An irony is that many of those same people would readily do the same via their PC or smartphone. They do not perceive such a simple-looking device in the same way they perceive the humanoid robot. The humanoid robot resembles a human. Their PC or smartphone doesn’t appear to be human-like. In their minds, this difference allows them to pour out their heart to the simple device, but they are hesitant or refuse to do so to the humanoid robot.
There is an entire research area within AI known as the uncanny valley that pertains to this human reaction when chatting with social robots and humanoid robots. See my detailed discussion about the uncanny valley at the link here.
Lots Of Mighty Ramifications
Once we have affordable and fully functional humanoid robots, the aspect of humans garnering their mental health advice via those robots has a lot of notable ramifications.
Consider these mind-bending questions:
- Will people be less likely or more likely to use generative AI for mental health advice if it comes from a humanoid robot versus using a conventional PC or smartphone?
- Will people opt to seek mental health guidance from humanoid robots rather than going to human therapists?
- Will human therapists essentially be replaced by humanoid robots, especially since people won’t have to go to a therapist’s office and can easily interact with the robots at home?
- Will therapist’s offices still exist, but be mainly occupied by humanoid robots as therapists, and have a meager handful of human therapists for those clients who still insist on human-to-human interaction?
These are all intriguing questions.
Layered on top of those questions are ones that are already being grappled with. For example, what is the legal liability associated with generative AI proffering mental health advice? What are the ethical implications? Is the privacy of those using such AI potentially compromised due to the AI retaining the interactions and possibly revealing those mental health interactions to others?
Those aspects will indubitably become much more visible to the public at large once there are humanoid robots taking on mental health discussions. Right now, the visibility of these concerns is generally low. You can expect much higher visibility when those walking and talking robots are doing likewise.
The Mental And The Physical
A final thought for now on this weighty matter.
Isaac Asimov famously formulated his several laws or rules of robotics and identified this especially noteworthy rule: “A robot may not injure a human being, or, through inaction, allow a human being to come to harm.”
We customarily interpret that line to apply to physical harm. What about mental harm? In the case of humanoid robots wandering here and there, dispensing mental health advice, perhaps we ought to widen the meaning of harm to encompass potential human mental harm.
Just something to earnestly noodle on.