AI is the flavor of the day. Billions are being invested to get needed research and energy.
This is probably good, but we forget that AI is an informational gathering tool and that as a tool it is dependent on past data.
AI can be especially useful as a partner with doctors, as it often can identify illnesses that a patient has through diagnoses.
But the human mind can make jumps in creativity way beyond data.
“History” showed that cotton seeds could be culled by slaves, but would AI have thought of removing cotton by using the cotton gin to separate cotton from seeds?
The Velcro innovation came about when George de Mestral noticed burs on his dog. He was able to connect how the burs attached to create Velcro (now few of us can tie shoe laces!)
I predicted the subprime crisis in late 2005 when the same data was available to everyone- banks loaning clients their downpayments, appraisals made as instructed, and refinanced loans to the poor. I looked at the data and made my prediction. But no one took me seriously.
Now please do not get me wrong. I simply love data which creates patterns, and AI is great at this, especially when you can extract results from that data, as with medical diagnoses.
Recently AI helped scholars to interpret ancient Latin; this allowed scholars to make connections, but there wasn’t an analysis.
Science News (10/25) described game theory where AI could be “programmed” to feel guilt. But, of course, AI can’t feel guilt.
But the danger with AI is that most humans are followers, and won’t use imagination to interpret data.
Carol Kranowitz’s “The Out-of-Sync Child” gives statistics on vision in humans. 80% of the information we take in comes from the eyes. About the same 80% is responsible for what we see. 93% of human communications is non-verbal. 75 to 90% of classroom learning depends on vision.
AI, of course, cannot “see”, another limitation. As an aside, those who work on AI might well wish to study how sightless humans compensate for their lack of vision. Learning their success here could enhance AI.
I like to tell the story of my friend Sam Jacobsen who operated a milk store in Madison, Wisconsin. Business was bad, very bad, so Sam decided to increase his hours from early morning to late evening. One day, a clerk mispriced his milk from 66 cents to 99 cents, and sales did not go down. Sam concluded that people would pay for convenience, and he built up a chain of convenience stores. Could AI make that type of decision. Not likely.
I do not want to be thought of as diminishing AI. It can be very useful. Robots will be taught manufacturing jobs and many service activities, for instance. But it is unlikely that AI will think like some humans do.
As a tool, AI is great as a calculator, a retriever of data, and a way to organize massive information. But I do not see the intuition or creative linking that some humans do for AI, yet!
One Response
Утепление играет ключевую задачу в удержании комфортной температуры внутри помещения.
Эффективная защита помогает заметно снизить расходы на отопление и охлаждение помещений.
https://rosmet-nn.ru/full/innovatsionnye-dobavki-v-kauchukovuyu-izolyatsiyu-povyshenie-ognestoykosti-i-antimikrobnykh-svoystv-1013/
Помимо этого, теплоизоляция препятствует появление влаги, плесени и разрушения стен.
Таким образом, качественная изоляция гарантирует энергоэффективность и комфортный микроклимат в помещении.