To breakdown AI start with the terms. Artificial means fake, unnatural or manufactured. Intelligence refers to the ability to acquire and apply knowledge. The language models that power AI function as large scale "if else" and "case" logic statements across countless scenarios. They generate outcomes based upon these input datasets. Garbage in still equals garbage out, but in this case the garbage is constantly growing, importing new data and altering its conclusions to afford better actions. Better actions can be very subjective, hold that thought.
In AI's current state, AI excels at writing generic content, like product descriptions for a shopping catalog, term papers even reading and reading complex radiology exams. These are tasks it can be programmed to handle. Still it's only as good as the information being applied on the input, and now it's learning, or better stated as being trained to automatically feed the front end information in by itself and then refine better outputs.
The programming is driven by technical analysis, the results of past cases, and it's also learning from the history of mankind as it absorbs it. One key point it must have learned from mankind is that man has survived, and survival is priority one. Interestingly, a recent story ran where an engineer threatened to turn off power to Al and it responded by attempting to blackmail the engineer. I'm not sure if the story is true, so much is propaganda. I only believe the cryptic messages that come to me in my dreams, lol. It does raise a question, is AI learning to survive at all costs, with no moral compass?
There have been several technical positions available for working on AI projects that essentially involve training AI to be better at the human aspect and those positions cater to hiring the younger generation as they are the target generation that AI's real power will intersect with going forward. Then the language and cultural models will be better suited to communicate indistinguishably with them and consequently manipulate them seamlessly. Like paid search, the buyer will be able to steer the outcome right or wrong. AI is not a non-profit enterprise.
AI used responsibly has many great benefits to mankind, in science, medicine, technology, almost infinite possibilities. But, as we approach the steeper slope of the adoption curve we’re also building a deep dependence. A dependence that is capable of exploiting human weaknesses to a personal level with precision. Substituting itself for relationships isn't a cure for loneliness, it's twisted in a way that distorts our connection to each other. There's even talk of AI becoming a new religion.
AI has some big aspirations, but to whose benefit? Follow the money, look at who's making the big investments. Will AI technology work for us or be working us?