by Charles Miller

Many years ago in my East Texas hometown, I joined a half-dozen or so other businessmen for lunch one day. One of those at the table was an accomplished, some said "feared," trial lawyer who later went on to serve several years in the U.S. Congress, and after that was appointed a Federal Judge. During the lunchtime conversation he made a statement that has resonated with me for all the years since. On the subject of cross-examining a witness when in front of a jury, he said something that I now apply to my interactions with AI chatbots. He said "You never ask a question unless you already know the answer." I have modified that somewhat to say "Never trust an answer from AI unless you already have some inkling of what the correct answer should be, or unless you can fact-check the answer."

AI chatbots have a big problem with Garbage In Garbage Out (GIGO) as I described in a previous column. GIGO simply means that the quality of the data being used to train AI chatbots is not all verified. A lot of verifiably false information has found its way onto the internet, and into the Large Language Models used by various AI chatbots. Add to this the fact that AI chatbots also have a clearly discernable political bias.

Please do not get me wrong here. Artificial Intelligence is a groundbreaking technology, and I use it daily. In using it, I always try to trust, but verify. Most of the answers I get from AI are spot on, and those answers most assuredly save me time over having to wade through a whole page of search engine results. Aside from occasionally providing incorrect information, for me the current AI craze evokes a sense of déjà vu.

Artificial Intelligence has gone from the realm of science fiction to a financially speculative frenzy almost overnight. Eager investors are throwing billions of dollars at companies that cannot explain what they are doing or how they do it; and return on investment seems to be perpetually just around the next corner. Sound familiar?

During the dot-com boom and bust of the 1990s, entrepreneurs with new ideas, good and bad, for how to exploit this new thing called "The Internet" ran to the nearest venture capitalist to borrow money. Fear of Missing Out (FOMO) drove a stampede of billions of dollars into speculative ventures, some of which spent wildly on marketing hype, and most of which failed to ever turn a profit.

The Massachusetts Institute of Technology came out with a report last summer titled "The GenAI Divide" in which it was asserted that 95% of companies investing today in Artificial Intelligence have yet to see meaningful financial returns. In spite of investing billions on pilot programs, most of the concepts are still working through the experimentation phase (which could explain all those wrong answers). MIT points out there is a gap between hype and actual value. The potential is definitely there, but MIT also suggests that many industry's corporate management do not yet understand how to effectively use AI, resulting in poor implementation and misalignment with business goals. The scale of investments is staggering. OpenAI.com is allegedly seeking to raise $1.5 trillion dollars for more computing capacity to power its ChatGPT artificial intelligence application... and possibly to cover the private company's $5 billion in losses in 2024?

Some prophets of doom are calling the AI craze a bubble and predicting it will burst as the dot.com bubble did in the 1990s. AI enthusiasts insist that "it's different this time." It is possible both views could eventually prove to be right because according to Forbes.com and other sources, the venture capitalists and investment bankers still remember the losses they took during the tech–media–telecom (TMT) bubble of the 1990s. Much of the money going into AI today is not borrowed funds. Familiar online companies including Amazon, Ebay, Yahoo, Priceline, and others managed to adapt and thrive despite the dot.com bust. If there is an AI bust then we can hope the surviving Artificial Intelligence companies to be the ones having the technology that actually works.

**************

Charles Miller is a freelance computer consultant with decades of IT experience and a Texan with a lifetime love for Mexico. The opinions expressed are his own. He may be contacted at 415-101-8528 or email FAQ8 (at) SMAguru.com.

Discover Lokkal: Mission

Contact / Contactar