All about A.I.

I was actually just reading about a law firm that was using ChatGTP to help prepare its legal briefs. In theory this is a pretty solid use of AI, as it can research a specific case law, all its rulings and legal precedents in a more thorough way and much faster than a human. The problem arose when the judge read the briefings and tried to manually research the case law history. It turns out that the AI had it's own opinion and fabricated fake court cases and legal precedent that would support it's conclusions rather than just using existing trials.

This also raises the question of who get held accountable for the bad actions of AI. There are already cases where an AI has created works using copyrighted material or said defamatory things about people in public spaces. But how do you bring a lawsuit software code?

Many years ago the stock market experienced what's called a "flash crash" where the market dropped something like 20%-30% in a matter of minutes. The reason was that there were large stock trading firms using software algorithms (not quite true AI yet) to look for buy and sell patterns in stock prices. Those firms using these programs had a competitive edge over the firms using only human traders. In general, a human can look at a lot of data for a company and get some good indicators of when would be good times to buy or sell that stock. However that one human can only be a subject matter expert in a limited number of companies, whereas an AI or algorithm can instantly get the information it needs to come up with the same buy/sell recommendations. What happened years ago was that these stock brokers had not only put these programs in charge of recommendations but also gave them authority to initiate trades. When a big sell signal flashed the software started liquidating assets which in turn created more sell signals. And because all the major brokers were using the same basic algorithms they all started liquidating their positions simultaneously. Now there are rules in place to limit how much control a program can have in such cases. Not AI but still a good example of how even well intended uses of this technology can go awry.
 
1685715344124.png
 
And here ya go...it will be taking over as your doctor if it hasn't already.