I have been hearing about “Artificial Intelligence” or “AI” for many years, but it seems that society is suddenly preoccupied with it. This recent craze about AI has been triggered by the introduction of human-like “ChatGPT” bots with AI in November 2022. Many companies are openly advertising “AI-empowered” products and services and discussions on AI are everywhere.
Most discussions also emphasize a fear of the unknown. Even Elon Musk, who successfully incorporated AI in the development of products like driverless cars has speculated that it is conceivable that AI can destroy human civilization by inducing a nuclear war or climate change or some other means.
Stephen Hawking, the renowned physicist reportedly made a similar comment years ago. The concept of AI in science fiction movies such as “2001 Space Odyssey”, “Star Wars”, “Terminator” and “AI” has been around for decades; but those were fictional.
Although AI was extensively used in industrial automation, the appearance of AI in our everyday life started with those automated menu-driven answering systems. As technology evolved, they could handle voice response, voice recognition and even facial recognition.
Technology was extended to virtual assistants like Alexa from Amazon and Siri from Apple who communicated almost like humans and had millions of pieces of information they could process in a fraction of a second.
Combining voice exchange with a mechanical task was the next step such as opening a door or turning on TV/light switches with a voice command. I was comfortable with these advances in technology. Occasional frustration resulted when the robots did not know how to answer some questions and a real person was unavailable.
Overall, I would probably trust a decision by AI more than that of a human being when it involved processing dozens of data and optimization. For example, what to do if I am suffering from some strange health ailments; it is obviously preferable to get just one opinion based on sound logic from AI than getting confused with opinions from multiple human doctors.
The same is true in getting investment advice or deciding the best place to buy a home or the best college to apply for. I may even be open to surgery performed by a robot. Although AI lacks knowledge gained by human experience and intuitive power, I cannot tell how much of a handicap that is.
I have not been in a driverless car but do enjoy various automated selfguided gadgets like Roomba vacuum cleaner, driverless train at an airport terminal or a coffee-serving robot. AI can reportedly write an essay or speech and even compose music.
My first discomfort with AI arose when things were being done not only without any input from me but also without my knowledge. It was spooky. My first such experience was encountering pop-up ads on my computer screen for products I recently searched on the internet.
A similar phenomenon happens on YouTube when it promotes videos based on what I watched recently. In both examples, I did not ask anyone to feed me with the information. I can even tolerate that, but it would get dangerous if AI is used to distort or even falsify information.
We live in the era of “fake news” and I do not believe everything that a news channel is broadcasting, but in those cases I know the biases of the news organizations and their anchors. Should I believe in AI? There are certain data sets I consider sacred: information about my bank balance and other financial information, my medical history, my personal data, etc.
I feel assured because organizations who have such data claim to have implemented adequate protection mechanisms: multiple passwords, encryption, etc. My main concern these days is if AI would have access to that information and even worse, ability to alter or steal the data. I am not paranoid about privacy.
My mindset is that I have nothing to hide and if everyone in the world finds out all about my conversations that is perfectly fine with me. A much greater concern is if AI can change election results or financial news to cause a major crisis in the country.
AI after all is not a machine; it is driven by software developed and run by real people with personal political, religious and other biases. This bias has already been revealed in companies like Meta, Twitter and even Google. A bipartisan bill is going through Congress which would ban AI from single-handedly launching nuclear weapons.
Recently I made a phone call to my bank, which is being bought out by another bank, to inquire about any change in the monthly service fee and other policies. I got the automated system offering me various options. As usual, I hit the “0” button which takes me to a live operator so that I can get my answers quickly.
However, after speaking to the lady at the other end for a couple of minutes, I realized that I was talking to a robot and not to a live person. I could tell because I talk to my Amazon Alexa all the time.
The giveaway is their emotionless impersonal monotonic voice with no touch of humour and no modulation in voice. I confronted her by saying that “I was hoping to talk to a live person and feel disappointed that I ended up talking to an AI like yourself ”.
She responded by saying “I am a real person”. This sent chills through my spine. AI has now been programmed to tell a lie. In addition, some features have been added to make her sound like a real person; one example was her making a “tu tu tu” sound while looking for some information just like a real person.
At one point, I asked her jokingly, “Hey, do you know Alexa and Siri?” That would have generated a chuckle in a real human being. She simply said “No, I do not know them”. She gave her name as “CJ”, reminding me of C3PO.
I am convinced that she was a humanlike “ChatGPT”. This was a harmless encounter, and I obtained the information I needed but I realized that ChatGPT is here; right at my doorstep and in my life disguised as a real person. Now I have all kinds of concerns.
The corporations are promoting AI for the sake of improving their productivity. The indirect manpower necessary to handle routine interactive jobs, such as secretarial jobs or customer service or making reservations can be drastically reduced.
However, it is now clear to me that these ChatGPTs can easily perform many other functions such as all teaching and training jobs, all agency jobs (insurance, banking, real estate, employment, travel etc.), all domestic chores, various engineering and architectural jobs and so on. Legal processes can be streamlined using AI.
Medical help can be dramatically improved with their involvement with fewer needs of health professionals. While many of these deployments of AIs can improve our lives, I immediately envision massive layoffs in various sectors resulting in possible economic devastations if not a complete change in our values and lifestyles.
Elon Musk might be right in predicting that AI will destroy human civilization but unlike destruction by a nuclear bomb or natural disaster that might happen gradually in front of our eyes and without us even realizing it.
(The writer, a physicist who worked in industry and academia, is a Bengali settled in America.)