AI Is Here – Is That Good or Not So Good
Artificial Intelligence or “AI”, surrounds us, it seems, at every step today. It might be thought of as an ever present but not always considered atmosphere in which we work and live – that is, in that way, like air. It has been criticized as replacing humans in some routinized professions but also praised as the wave of the future that will cure illnesses, solve seemingly unsolvable problems, improve our lifestyle but also, sometimes, rob us of personal connections. It makes everything easier – or does it?
So, when I recently signed into a Zoom meeting (another modern convenience) I noticed that AI “signed in” also. AI was dutifully summarizing the meeting typing along in whatever mode it had been trained (although failing to state that I was perfectly capable of doing this myself) and I did notice, of course, at the conclusion of the meeting that it attempted to reflect accurately what was stated. It probably typed a little faster than I do but generally all went well. Here is where something odd occurred and here my human mind had an answer.
AI stated in its summary that I and “Samsung” did something together. In all honesty I know no one named Samsung” and “Samsung” did not participate in the meeting. So, where did AI, the Zoom guru, get the idea that Samsung and I met?
The answer is simple and this brought me back to one of the statements made in one of the many AI training sessions I attend as an attorney while learning our new billing/scheduling software. (As an aside, if I appear in the past few months to be absent to some of my friends and unfortunately once in awhile to a few of my clients it may be due partially to attending multiple instruction sessions in all of the great new features that are going to make life more convenient while practicing law by using AI in my practice.)
To return to the question, the statement I remembered was “if AI cannot recognize or answer a problem ‘it makes something up.’” This answer makes great sense to me as a former high school social studies teacher and a former adjunct professor at university. Students do this all the time and now an AI program was seemingly very human and acting likewise. AI could not identify the other participant in the chat and so “made up” the closest name it could think up. That name is “Samsung.” So “Samsung” attended the meeting.
I describe this simple experience for a reason. AI can be fast. It can even be accurate but it can trip over lots of situations that humans would recognize as mistaken. I think of jobs, some of them held by a number of my clients that require numerous items to be described and sorted. Of course AI can do it faster. Will it do it right? Sometimes, probably most times it will, and reprogramming over time with the assistance of humans can and probably will make it better. Still, there are areas where human experience and intervention are needed.
I also think about healthcare – an area, along with estate planning and estate administration, where I spend a great deal of my time. I think, for instance, of decisions regarding whether treatment for a given medical condition is to be continued or discontinued and whether it is to be covered by one of many health insurance plans available to the individual patient. For some patients under some plans algorithms have been chosen by insurance companies to decide whether an individual should receive or should continue to receive original or followup treatment. It would seem that some decisions might be so connected to the individual and so dependent on human judgment that AI might not be capable of making the “correct” decision. Financial considerations regarding the person and the plan can also bias a decision when fed into AI.
On the other hand massive amounts of clinical data over time might suggest new solutions for care in the individual case.
So it seems there is no absolute answer whether AI is “good” or “not so much.” I clued that answer into AI itself and found that according to itself it is “good” for efficiency and productivity. Also for scientific breakthroughs, accessibility and advanced analysis. It is “not so much” in “lack of truth” (hallucinations), bias, skill erosion and environmental and economic impact.
About the Author Janet Colliton
Esquire, Colliton Law Associates, P.C. Janet Colliton has practiced law for over 38 years, 37 of them in Chester County, Pennsylvania, a suburb of Philadelphia. Her practice, Colliton Law Associates, PC, is limited to elder law, Medicaid, including advice, applications and appeals, and other benefits planning including Veterans benefits, life care and special needs planning, guardianships, retirement, and estate planning and administration.