Artificial Intelligence is perhaps the most promising technology for transforming our lives — but it's also incredibly scary. At CES 2022, A panel of AI experts discussed what role AI might play in the future of healthcare.
Can you trust AI?
In a session titled "Consumer Safety Driven by AI," Pat Baird, an AI engineer and software developer who works in standards and regulations for Phillips, and Joseph Murphy, VP Marketing at Sensory Inc., an American technology company that develops AI products, discussed what AI could add to our lives. They also discussed the apprehension many people feel about the technology.
Baird said that getting people to trust AI is essential to overcome those misgivings. He said that many find the technology weird and wonder who designed it. Consequently, they're uncomfortable receiving care from an AI — and will likely continue to feel that way unless they can be convinced that it's a positive invention.
A lot of this skepticism is rooted in myths about AI. For the tech to be widely adopted, it needs to build up goodwill among users and regulators.
But it's not a matter of building blind faith in AI. Baird warned that "overtrust” happens when people believe the technology too much and “stop doing critical thinking.” For example, many accidents in self-driving cars happen when human riders stop paying attention, start playing on their phone, or engage in some other distraction. That left them unable to take control of the car before it crashed.
For Murphy, AI is something to be excited about. There are already millions of AI experiences every day, and most of them work well. Those that don't are the AI stories that often get attention. He brought up the recent example of a ten-year-old girl who was told by Amazon’s Alexa to stick a penny into a live socket as part of a challenge she asked the AI to come up with. The dangerous suggestion got a lot of media attention but as Murphy says, it was a situation where the AI still needed oversight and a human applying common sense, something that in his opinion would have to remain part of the process of living and working with AI.
Baird felt that AIs are likely to perpetuate inequalities that already exist in society. Could AI improve those problems instead? He argues that software reflects the values and biases of the people who create it, making it a mirror of society. Murphy, who works with speech recognition, acknowledged built-in biases that affect his industry but also believes that the “industry as a whole is lifting”. It comes down to training the data in a way that truly represents the population, he said.
How can we improve AI safety?
Community sharing of white papers and ongoing conventions help improve the technology, prevent misuse, and spread best practices among the speech recognition industry, explained Murphy.
AI could also be used to improve the safety of the humans who use it. The most exciting tech development happening right now according to Murphy is the emerging metaverse, with people existing in an increasingly realistic way in the virtual space. He believes it will be used for training in hazardous environments, among other examles, as well as for hobbies. An electrician, a first responder or a surgeon could perfect their techniques via artificial intelligence, before trying dangerous procedures in the field.
“Before we get machines to talk to each other, we need to be better at getting people to talk to each another, “ also shared Baird, pointing out that machine learning will become better once people improve their “collective learning”.
AI in healthcare and beyond
Baird believes in the healthcare industry, improving the efficiency of the doctors could be improved greatly through AI.
According to him, the United States currently has approximately 1 physician per every 300 people. AI could achieve an improvement in the efficiency of administering healthcare for such a large group that could be very impactful. The technology could help take over busy work and give the doctor more time to focus on the patient.
Baird sees that some may see machines as a dehumanizing feature of healthcare, but he argues that machines could actually provide more direct time between doctors and patients. Direct interaction with AI could make it easier for people to talk about sensitive issues they wouldn't feel comfortable discussing with a human doctor.
Education technology is about to blast off with AI, claimed Murphy. That’s also where the trust of AI technology is the highest. Another area of greatest AI impact is likely to be autonomous driving and autonomous features in vehicles, like notifications of environmental factors.