Disclaimer: the predictions in this post are based on my opinion and are not all research-based or backed.
Many jobs that humans have performed have or are being replaced by some type of technology. Some examples include:
- Switchboard operators, telemarketers -> Computers
- Receptionists -> Check-in kiosks
- Bowling alley pin resetters -> Robots
- Bridge toll collectors -> Automated toll lanes
- Check-out cashiers -> Self-check-out lines
- Ticket sellers and airplane service desks -> Computer kiosks
- Assembly-line and factory workers -> Automation robots
- Bank tellers -> ATMs
- Warehouse movers -> Amazon fulfillment center robots
- Fast-food workers, chefs, cooks -> Automated robots
- Pharmacy technicians -> Robots
- Information gathering, analysis, and research -> Cloud computing and A.I.
- Pilots -> Drones
- Bomb Squads -> Disposable robots
- Journalism and writing -> A.I., e.g., OpenAI’s GPT
- Typist -> Computers
- Stock traders -> Computers
- Postal workers -> Sorting machines and robots
- Room service -> Robots that can deliver items
- Surgeons, Doctors, Techs -> Robot-assisted surgeries
The Technological Race
Humanity is competing to develop the best and smartest A.I. This documentary shows some insight into how the West and the East are competing to be the leaders in A.I. technology. Any technology of course can be used for good and bad. In the post-Snowden era, recent advances in technology have raised privacy concerns that resemble the fears of the book 1984. Think about how technology has changed and made its way into your life.
Technology in our Lives
My home now has Amazon Alexa devices in two living spaces with microphones that can listen in when alerted to the wake word. My door locks, garage door, lights, HVAC system, even my bedroom can be turned on and off with a voice command, from my smartphone, or from certain pre-programmed algorithms that run on the cloud. A robot litter box automatically detects my cat, empties the litter into a waste receptacle, and alerts my smartphone if the bin is full. These “Internet of Things”, or IoT devices have made their ways into our homes and lives much faster than I first predicted. But is all this technology A.I.? Not quite. They can do perform simple tasks based on a “if this happens” and “then perform such action” premise. I would consider it “internet-connected” technology. However, it will only get better and smarter.
How does the future of A.I. look? As of 2020, in China, citizens are already paying for items with facial recognition (which relies on A.I.) – no need for even a smartphone.
Can A.I. be Creative?
Behind the scenes, cloud computing, A.I., and deep learning are advancing at a rapid rate.
The documentary shows that A.I. has the potential to solve problems by offering creative solutions.
That is one beauty of AI. Two examples from the documentary show how machine learning and A.I. defeated a world champion in the board game Go and exploited a tactic in the video game Brick Breakout to win by doing the unthinkable.
So A.I. can be creative and solve problems but is it a threat to occupational therapy? Occupational therapy requires a complex skill set of knowledge, practical skills, and more importantly – creativity and problem-solving.
Will Robots Take My OT Job?
The website willrobotstakemyjob.com predicts a 0.4% chance of robots taking over occupational therapists’ job.
I strongly encourage you to check out this website which is based on Carl Benedikt Frey & Michael Osborne’s “estimate the probability of computerisation for 702 detailed occupations, using a Gaussian process classifier” combined with data from the BLS (which we are big fans of).
For comparison, we plugged in “fast food cook” and the website returned a whopping 81% chance of the job being replaced by robots. This is likely to be inevitable as this industry depends on efficiency for success.
It’s safe to say that A.I. won’t take OT jobs anytime soon. Here’s what I think.
A.I.’s Threat to OT
From the documentary shared above, we learned about the potential for to far exceed the human cognitive capacity for comparing multiple outcomes and offering creative solutions – something that occupational therapists thought A.I. could not replace. The real threat comes when you take a top-down approach when you can envision a robot with sophisticated enough A.I. to do all the things that an occupational therapist can do. Even if a robot can do some, such as half of the job that an OT can do, COTAs and OTs themselves would still be at risk.
Let’s take a bottom-up approach to divide an occupational therapist’s skillset into several categories:
Recent advances in technology can make A.I. write very much like a human. Take a recent GPT-3 social experiment for example. A college student’s fake, AI-generated blog with GPT-3 fooled tens of thousands – and even made it on a top list. If such a technology can write such convincing articles, it could likely handle the documentation that occupational therapists do. After all, most of the documentation can be quite repetitive and predictable with minimal creative writing involved. Take for example how you write a patient transferred from bed to a commode. Since A.I. would perform these tasks, it could be programmed to report and document these tasks even more accurately for the level of assist (based on weight sensors) for min/mod/max assist down to the percentage. If standardized across robots, it would be far more sensitive and reliable than a human OT’s estimate for assist levels.
Documentation: A.I. wins
Robots have gotten a lot better at communication. However, to replace OTs, they will need to talk to patients, doctors (robot-doctors?), family members, answer e-mails, phone calls, read and act on schedules, and so on. Robots will also need to be able to handle diversity in culture and multiple languages (without the use of interpreters). Look how far Amazon Alexa can answer your questions. While it is very good at that, it cannot have a conversation with you – yet.
Communication – humans win
This is a complicated set of skills that also demands interaction with the environment. Examples include navigating from offices, gyms, hallways, bathrooms, showers, outside terrain, climbing stairs – all while potentially assisting a patient safely to prevent injury or falls. My robot vacuum struggles with dog toys, different terrains, and certainly cannot handle stairs. This technology will get better but is not even close to navigating dynamic and challenging environments. Pay attention to the self-driving cars and truck industry to see how this goes.
Navigation: humans win
Knowledge and Memory
Computers already have more knowledge than humans. Databases such as Wikipedia is an example of this (although one could argue that this is user-driven). Computers can gather knowledge much faster than a human, who would go and do a search on a search engine anyways. In the time that an OT would do a search, A.I. could search through, analyze, and make an informed decision better than a human. All the things that you “learn” in OT school could be “learned” by A.I. As for memory, computers have a far superior memory capacity than humans for “storage”.
Knowledge & memory: A.I. wins
Work Experience & Clinical Reasoning
You’re probably thinking, okay, but A.I. does not have the experience that I have. That may be true as you worked with a patient and gain experience on many skills, tricks, and techniques for many years compared to a fresh robot. But this is something that A.I. could quickly learn as well – if not faster. Nowadays, data is king. That is why companies like Amazon and Google are on so valued – they are constantly analyzing behaviors and learning them to make predictions and recommendations. Netflix video recommendations is another example. An OT-robot that is cloud-connected could gather the experience from multiple other sessions, remember them, and learn from them. Still, humans have the edge when it comes to clinical reasoning.
Work Experience & Clinical reasoning: humans win
OT’s do a lot of transfers. However, we are limited by a certain extent in strength and rely on machines to assist. Take ceiling lifts, hoyer lifts, slippery sheets, hover mats, adjustable hospital beds, even wheelchairs for example. They currently help to do the jobs of our transfers. Can robots assist completely for a transfer without the help of an OT?
To assist a patient with a transfer, the robot would need a lot of finese, be gentle and sensitive enough to human’s when they experience discomfort or pain, and act accordingly for safety to prevent injury and falls.
Believe it or not, something like picking up various shapes and sizes can only be done by humans at Amazon fulfillment centers. Amazon is currently investing a lot of money to solve the problem of packaging various items into boxes for their fulfillment centers. If they solve this problem soon, such technology would result in a lot of lost jobs at Amazon and could be improved and adapted for the healthcare’s patient care industry – but not anytime soon. Another barrier would be human adoption. Older generations would likely be resistant to having a robot assist you with transfers, but younger generations may embrace such a technology in the future.
Transfers – humans win
Assisting with ADLs
Something like donning/doffing clothes can be very complicated for a robot to do. Humans have different body make-up and robots will need to provide the just-right challenge. Could a robot help assist tying a hospital gown drawstring around a patient who is wearing a cervical collar? This is a very complicated task that I don’t see robots doing anytime soon.
Can a robot assist with bathing body parts are with toilet hygiene? This is a very private and delicate task that humans can only do safely. Another issue is again the barrier to adoption of having a robot in your shower help you.
Many would have privacy concerns of cameras and microphones recording your intimate encounter as A.I. would need to learn from data, and this means to some extend recording activities such showers. For this reason alone, many would oppose robots assisting with toileting and bathing and therefore advancements in robots assisting with ADLs.
Assisting with ADLs – humans win
Having a robots act and behave like a human is a little freaky. However, another approach may be to have robots be more approachable by acting cute, such as an animal or other character – instead of a human. Currently, we interface with technology with screens or in the background with voice-assistants. In the near future, interactions with A.I. will be more up close and personal. Robot vacuums and drones are starting to be more prevalent in our society.
Human-Like Interaction – humans win
Robot-OT vs. Human-OT Summary
- Documentation: A.I. wins
- Communication: humans win
- Navigation: humans win
- Knowledge & memory: A.I. wins
- Work Experience & Clinical reasoning: humans win
- Transfers: humans win
- Assisting with ADLs: humans win
- Human-Like Interaction: humans win
AI: 2, Humans: 6
This is of course not an exhaustive list of the OT skillset that robots will need to replace. For A.I. to replace OT jobs, there are many things that need to be developed, tested, and proven to work before adoption. A recent example of disruptive technology is Telehealth. We are still far from widely adopting such a system even amidst the COVID-19 pandemic.
- Ethics – OTs encounter ethical issues on the job, how would a robot act? Would it be the same every time? Could A.I. even recognize and consider ethics for real-life scenarios and problems?
- Privacy & security – we mentioned some privacy issues concerning ADLs with toileting and showers. What about everything else? Robots will collect and analyze a lot of data. This is big money. Hackers will want (and likely be able to) gain access to this data. Imagine if a hacker was able to see your OT session during your most vulnerable moments as a patient. Robots will need to be secured, and this can only be accomplished through improvements in the technology – but is never foolproof. Anything can be potentially hackable.
- Data – who owns the data? Can the data be bought out or shared with third-parties? To what extent do patients have control over such data? Who would be held liable if such data was compromised?
- Liability – who would be held liable for mistakes, injuries, even death? The robot? The company that owns it? This is new territory that we have not encountered before.
- Insurance – would insurance cover something but not others? Would robots be made equal in quality? There will always be tiers and different levels and quality of products – in any industry. What if a hospital and insurance pay for an inferior robot or technology? Would you be getting the best care?
- Decisions – what happens if a robot disagrees with a patient or family member? How will the robot act?
- Collaboration – OTs collaborate with many disciplines. Robots will need to integrate with all of these disciplines to a certain extent – even if it still involves a human-held job. This likely will involve training for the other disciplines to interact with an “OT-Robot”.
- Advocacy – would a robot advocate for their patient? How would you know that it is acting in the patient’s best interest and not other parties’?
- Back-up for failure and power – what happens during a blackout? Will the robot run on battery? How long can it last? What happens if the Internet goes down? What happens if the robot breaks mid-session? Who will service it?
- Maintenance – how often will robots need to be serviced? The costs will surely be high. Will companies neglect to service their equipment?
Potential for A.I. in OT
While we mostly discussed whether A.I. and robots will replace OT, it is worthwhile to consider how A.I. can improve OT and make it a valuable profession.
Research & Education – What do OT’s mostly do? Teach! A.I. is already assisting researchers with analysis, finding patterns, simulations, and so on. If OT is to not be left behind, it should harness the power of A.I. to assist with research to provide the best and creative treatments for patients. After all, technology will be more integral in our future lives, so it is natural to integrate this big data in the research, but with an OT perspective. A.I. can also teach practitioners, students, new grads the best practice based on previous data. We are already using data (such as database searches, systematic reviews) in research. If we leverage the power of A.I., it can creatively analyze certain and multiple datapoints we never even thought of to predict with high probably better patient outcomes. Very exciting stuff.
Prevention – has been a hot topic in healthcare as it can help to reduce costs and is part of recent healthcare reform and vision. Advancements and early detection are sure to be made with A.I. for diagnostic imaging and research of diseases such as breast cancer. OT will need to play a role with working with A.I. in detecting higher risks for fall, underlying conditions, even cognitive decline – and provide the just needed assistance to meet the patient’s cognitive deficits. How many patients do you know who have poor memory, judgment, safety awareness, planning, problem-solving? If A.I. can detect declines in such cognitive function and even report to clinicians, it is a win for healthcare.
Assistive Technology – I can see this space being even more accessible and innovative for OTs, especially in the aging in place space. IoT devices are already being accepted in our homes. Many systems can be developed to help older adults age in place safely and to help provide reassurance to family members of their safety. Imagine A.I. that reliably detects that patients have a fall in their home (with multiple sensors) and automatically alerts EMS and family members (with provided permission, of course). We are not that far off as the Apple Watch can detect falls already. OTs have long valued assistive technology and should be more involved with A.I. to stay relevant as a profession.
3D Printing – hopefully, will be more prevalent in OT. 3D printing has become more popular and it has been used during the pandemic to make mask comfort ear-guards. Imagine being able to print a piece of adaptive equipment on the spot for your patient and being able to bill it to insurance. In a futuristic OT world, A.I. can help make recommendations based on body scans, draw up a design, and print the design for you, e.g., a custom hand splint.
Mechanical Assistance – A.I. can assist with transfers not just for practitioners working with patients, but imagine them being purchased to be provided in the home permanently. One major barrier for aging in place is the environment. A.I. can help us overcome our environment. Multilevel homes with stairs can be overcome with more sophisticated A.I. and robots that are even more advanced than stairchairs. Self-driving cars are sure to take off soon. We will see how this plays out for OT driver rehab. Since driving in society is valued because it is associated with independence and freedom for many, A.I. will be a great disruptor to OT and community mobility – and OTs will need to keep up.
As you can see, there are many obstacles to overcome if A.I. and robots are to take the jobs of occupational therapists. Technology can improve the way we live our lives. In the near future, I see A.I. not replacing, but working along occupational therapists to make the experience better for both patients themselves as well as on the job. OTs should work closely with experts to develop and test new technologies that integrate A.I. into the forefront and backend of healthcare to continue to stay relevant and not be left behind.