By Dr. William Oliver Hedgepeth
Faculty Member, Transportation and Logistics Management

and Dr. Wanda Curlee
Program Director, Business Administration

Humanoid robots have been around since 1495. That was when Leonardo Da Vinci took a suit of armor and made it capable of sitting and walking, just as it would do if a human were inside it.

Da Vinci’s “robotic knight” has not survived, and no one knows exactly what it was capable of doing. But apparently it was driven by a system of pulleys and gears so it could walk, sit down, and even work its jaw.

Today, robots armed with artificial intelligence (AI) are being used to mimic human intelligence in many fields. AI properly belongs to the field of computer science.

Machine learning, another term for AI, solves a range of problems. Robots use AI machine learning aspects and data to mimic human movements and solve problems, including surgery. And robots are especially helpful performing nursing operations and medical assistance during the coronavirus pandemic.

Meet Sophia, the Humanoid AI Robot         

David Hanson, through his company Hanson Robotics, created Sophia, a humanoid robot powered by AI. She has facial expressions, cracks jokes, holds eye contact, understands human speech and recognizes faces, among other things.

The United Nations made her the first global citizen, and she is the first robot named the Innovation Ambassador for the United Nations Development Program. She was also made a citizen of Saudi Arabia in 2017, the first robot in the world to achieve that status.

Sophia has broad cognitive learning capabilities. As Hanson explained, “We feel that for realistic robots to be appealing to people, robots must attain some level of integrated social responsivity and aesthetic refinement.” “Rendering the social human in all possible detail can help us to better understand social intelligence, both scientifically and artistically,” he added.

Sophia has arms and has rudimentary legs. As robot limbs technology progresses, there may be a day when Sophia – who was modeled to look like actress Audrey Hepburn – may not be recognized as a humanoid robot except by another humanoid robot.

Robotic Surgery’s Positive Side Effects

According to Benjamin Tee, robots in the operating room mimic the hands of surgeons. Robotic surgery reportedly leaves less scar tissue and fewer scars than a human surgeon. The pain and blood loss are minimized, too. Moreover, patient recovery time is said to be quicker, requiring fewer days in the hospital.

Key elements of the humanoid robot are its skin and touch. The TEE Research Group is working to improve their neuromorphic robotic skin, a term introduced by famed Caltech engineer Carver Mead. This robotic skin is modeled on our own skin. How the robot holds a plastic bottle or a surgical instrument is based on AI algorithms and massive reams of data about what it means to touch and have a vision of what that touch is all about.

Picking up a soft drink can is easy for us to do. But the robot needs to figure out the shape of the can, for instance.

And since cans come in different shapes and sizes, different data is part of the robot’s AI decision-making process. The robot must have data to decide on the proper amount of strength to use with those mechanical fingers.

But for surgical applications, robots must have more than just a vision of what to hold. They need to be able to touch parts inside the human body, which are not always visible.

The National University of Singapore is creating a robot AI sensory system, or AI brain, which is supposed to operate on data like our own neural network system. The AI brain and the robotic fingers, combined with artificial skin and vision sensing systems, will be able to make possible surgical decisions.

Moving Outside the Operating Room

According to Analytics Insight, AI humanoid robots are now being used to help children and the elderly. Children undergoing painful or scary hospital procedures have found some solace in human-looking AI robots. They capture the child’s attention through funny stories or other means depending on the child’s age.

Analytics Insight points put that a “most interesting implementation of humanoid is Autism Special Education. The humanoid robot can engage children with applications designed for special education.

“The NAO robot is especially suited to interact with autistic children because the NAO robot is interactive and fun, engaging and captivating, and adaptive to the needs of the classroom from individuals to groups. It is a great help for teachers that appreciate eliminating monotonous tasks.”

Robots are also helping the elderly in many ways. One of the newer inventions is AI robotic arms that have tactile functions. These arms were designed and tested to move patients from a bed to a wheelchair and back to the bed. The arms learned how much pressure to apply depending on the weight of the patient.

Is There Bias in the AI Brain?

AI and robotic systems rely on vast and complex streams of data. But just how accurate is the data in an AI or robotic surgical system operating on patients? Is there data bias? 

When it comes to the worldwide pandemic, how data is collected might involve unknowingly creating bias in some manner. That bias may inadvertently be used to determine how patients are selected for trials or surgical procedures.

Diversity, too, is important. In a homogeneous county where most citizens are one race, one color or one religion, diversity might be viewed differently than the data collected in the heterogeneous United States. Combining this in a strategic or global database might produce misleading data in terms of medical treatment cause and effect.

A question then would be how do data analysts collecting this COVID-19 data focus on a strategic segment of our diverse society? Or does that AI model focus on a specific population or segment of society to estimate some aspect of cause and effect?

How Many Mistakes Are Made?

Besides data, AI robotic systems must be trained continuously to evaluate the successes and failures of medical procedures. How many failures will a surgeon or a patient tolerate? Do patients know that some AI or robotic systems are assessing them or assisting the surgeon?

How much loss of statistical accuracy from an AI system will doctors accept? How many years does an AI robotic database go back? When considering billions of pieces of data on medical procedures that worked or did not work, does that matter to specific patients? Is data more than five years old too old for an AI or robotic system to be trained on?

How do you define and then measure AI and robotic accuracy and failure? Is there a difference between unfairness and unbiased data? Is there a difference in fairness and no bias?

Robots Are Not Everywhere

While humanoid robots are seemingly sprouting up everywhere from elementary schools to nursing homes with great, almost human-like affection, they may not remain so for long. For instance, this pandemic has had a negative impact on Walmart robots.

According to the Wall Street Journal, “Walmart Inc. has ended its effort to use roving robots in store aisles to keep track of its inventory, reversing a years long push to automate the task with the hulking machines after finding during the coronavirus pandemic that humans can help get similar results.”

This pandemic has created a new normal whereby Walmart employees pick out products for customers waiting in their cars outside. Employees who spend their days filling shopping carts or baskets for online customers have become inventory managers now. These humans see the inventory is being depleted.

Robots were tracking inventory for Walmart, but with the advent of employees fulfilling the orders the tasks were duplicated. Human touch is needed. Walmart knew this, so they kept humans and have diverted the robots to other tasks.

But do not discount robots totally from Walmart’s and other retailers’ business and technology decision-making. Humanoid AI robots are here to stay.

About the Authors

Dr. Oliver Hedgepeth is a full-time professor. He was program director of three academic programs: Reverse Logistics Management, Transportation and Logistics Management and Government Contracting. He was Chair of the Logistics Department at the University of Alaska Anchorage. Dr. Hedgepeth was the founding Director of the Army’s Artificial Intelligence Center for Logistics from 1985 to 1990, Fort Lee, Virginia. 

Dr. Wanda Curlee is the Program Director for Business Administration. She has a Master in Technology Management and a Doctor of Management in organizational leadership.

Dr. Curlee has been teaching online for over 20 years. She currently researches Artificial Intelligence topics. Dr. Curlee is active with the Project Management Institute (PMI). She currently serves on PMI’s Ethics Review Committee. She has several certifications with the Project Management Institute.