This is the second part of a two-part blog series on the disruption of AI. Catch up on Part I here.
AI Forecast
So, what should we expect from AI…or more specifically from open AI platforms like ChatGPT? An article looking at the effects of AI on the economy appeared a few months ago in CNN Business that predicted global job losses could be as high as 300 million. They predicted the losses would be focused on white-collar jobs.
OpenAI, the company that developed and released ChatGPT, runs on Azure, Microsoft’s cloud development platform. This means that ChatGPT is not just open-source code but an easily available platform that allows any company or individual to develop AI applications or just leverage a fully functioning AI engine. Anyone can register with OpenAI and log in to ChatGPT. You can use the ChatGPT platform to ask questions like, explain quantum computing. Microsoft invested $10B in OpenAI this year in the hope of using ChatGPT to challenge the ubiquitous Google engine.
What Jobs Could Be at Risk?
When new computing capabilities emerge, it usually puts jobs at risk. For instance, when the first engineering design applications came out, draftsmen lost their jobs because the engineers could render their own images with the platform. As the engineering applications became more sophisticated, more and better work could be done with less engineers, letting the computer do the difficult math and reducing human error. The same will be true with ChatGPT. The open-source AI platform can transform dialog into images. It can solve complex analytical problems. It can use a database of diagnostic data to better analyze a patient’s condition and recommend therapies. According to a CNN Business article, the AI system was 21% more accurate and 41% more empathetic to patients than their doctors.
Any job where analysis is critical could be in jeopardy. Think about insurance actuaries, running complex problems to determine if groups of individuals like; OBGYN doctors, Florida homeowners or police officers should be insured. By using AI tools like ChatGPT, analyses could involve more data about the individual, like shopping habits from credit card company partnerships and video surveillance using facial recognition. That might indicate that the individual has very healthy habits, like running or walking, or unhealthy habits, like excessive drinking. This data combined with the analysis of claims for drunk driving accidents could help better analyze risk resulting in higher profits through less liability. What about Merger and Acquisition departments? They need to comb through reams of data to determine if companies can be more profitable when combined. That would be easy for AI.
Microsoft and AI
I would imagine we can expect tools like Microsoft Excel to incorporate ChatGPT into advanced versions. That would keep Microsoft at the top of the spreadsheet food chain. Microsoft first invested $1B in OpenAI in 2019 and I have experienced advancements in the capabilities of Microsoft Word, haven’t you? Word is much better at recognizing where commas belong and suggesting alternative phrases and the reduction of excessive wording. Google AI gets better at anticipating what I am really looking for and serves up more appropriate matches.
Potential Consequences for Business
Businesses should be looking for a way to incorporate AI. If your business is focused on manual labor like window washing or landscaping, you are probably safe for now. Computers cannot fix plumbing, pull weeds, put down mulch or install electrical wiring. White-collar jobs are different, especially if they are problem-solving jobs.
The future looks promising for the application of ChatGPT to solve big-time problems like cancer. Loading in tons of data and looking for discrete patterns should be no problem (except horsepower). The bad news is something that OpenAI talks about specifically on their website. They are concerned about the ethical issues posed by releasing this kind of computing power to society. They left me with the impression that there was much more capability available than they have released.
OpenAI wants to make sure that there are societal safeguards in place as capabilities are released. They indicate that the power will be difficult if not impossible to contain. They are giving good people the ability to do great things but they are also giving the not-so-good people tremendous power. Here is an example…
A couple months ago my American Express card was compromised. I got an email from Tesla, telling me my new charger was on its way, but I didn’t order a Tesla charger. I checked my Tesla account and sure enough, there was a charger on its way to New Jersey. I called American Express and let them know that I did not order the Tesla charger. They put me through to the fraud hotline – this fraud scheme was something they had not seen before. The criminals had figured out my future American Express account number, one which I would not have for years. Apparently, there is a pattern to the account numbers so the criminals (undoubtedly, their AI platform) chose a number that was at least ten years out. I didn’t get a notification from American Express since it was not an active number.
Companies that are susceptible to AI crimes need to implement AI to proactively manage fraud. I imagine American Express has plugged the hole I identified, but what will the criminals think of next?
Who Drives a Tesla?
One way to see AI at work is in your Tesla. Tesla’s Autopilot AI team has been hard at work for years refining the capabilities that make driving a breeze for me. This is my third year with my beloved Ethel, experiencing the genius of the AI team. For the first two years, self driving was about staying in the lane and managing cruise control at a safe distance (that you decide) from the car in front of you, exiting off freeways, changing lanes (with permission from the driver), letting you know where you had to stop to charge up while driving from Birmingham, Michigan to Petoskey, Michigan, etc. Now, Ethel can make complete turns, safely change lanes and navigate around a crowded city!
It is like an amusement park ride! My poor husband is not a fan of the ride but I love it! It is a stunning example of the brilliant AI team. I have experienced how the team learned about an issue and then in the next software update Ethel improved. A couple of years ago, Ethel and I would be driving along on the freeway in the right lane and a car would drive down the on-ramp to merge into traffic and Ethel would spot the intruder out of her right side camera and FREAK OUT. She would slam on the brakes and scare me to death. This might be one of the reasons that my husband is not a fan of the ride! This was clearly reported to the AI team and in a future software upgrade it was fixed. Now Ethel notices the merging driver and simply slows down to allow them to safely enter the motorway. I was so proud of her the first time she did it!
She is always learning. Last week, Ethel and I were driving home (heading west) and she had to make a right turn (going north) and immediately merge across 4 lanes of traffic, get into a boulevard left turn (if you are not from Michigan, that is a u-turn around a landscaped median), which means she had to go sideways west to make the u-turn to head south. As she made the right turn, she started to head across the lanes and got past the first two but then ran into traffic (people who were more familiar with the boulevard concept than Ethel) and couldn’t figure out what to do, so she just headed north while she figured out how to do the u-turn. In less than a minute, she had reconfigured the map to show me where she was going to make the u-turn to head south. Unfortunately for Ethel, it was a couple miles up the road, passing several turn-arounds that would have been much quicker so I took over and made the turn for us. Ethel and I are a team after all.
I am confident that over-time Ethel will be a great driver. I am counting on it because I want her to drive me around when my children tell me that I am too old to drive. My Mom is 83 and still has her car but we only “allow” her to drive across the street to Kroger to get cookies and ice cream. She has a Verizon Hum plugged into her 13 year old GMC and any time she pulls out of the garage, the Hum sends me an email and a text message to alert me so I can make sure she doesn’t get lost. When I am 83, I will want to go farther than across the street to Kroger so I need to help Ethel so eventually she can free me. I would love to say that my Mom should trade in her GMC for a Tesla but alas, the amusement park ride might be too much for her. She grew up with a manual transmission and no power steering, Ethel would seem like a spaceship for her. But I am going to work with Ethel and her AI team so that someday maybe Ethel and I can fly.
What would you want to solve if you had an AI engine to help you think?
