AI: still learning its own game

By  Jeffrey Tomas Marchesseault

We see and hear about it all the time. Artificial intelligence is the world’s latest, greatest business technology standard because it speeds up research and development, automates many of the time-consuming tasks traditionally performed by human employees, saves money on processes and personnel, and helps grow revenues and net profits faster through accelerated customer conversion and an extremely efficient allocation of resources.

Sounds about as worrisome as it does promising. The globe-sweeping AI revolution is gratifying for those who are tech-savvy enough to jump in with both feet kicking, disconcerting for businesses that don’t know where to begin, and alarming for any employee who doesn’t know whether they can upskill fast enough to save their jobs or change careers.

While taking a deep gasp and preparing for the sudden impact of earth-shattering change may seem like a terrifying ultimatum, most of us can still take comfort in recalling that new technologies almost always integrate their way into our lives gradually in almost unnoticeable ways.

Early adopters on the bleeding edge of consumerism get a head start by lapping up the less refined and less affordable versions and prototypes of the more refined and affordable gizmos and gadgets that the rest of the market eventually adopts. Improvements are constant until we reach a plateau where most software and hardware is commodified into a new standard of performance that was unthinkable ten years ago.

If you’re a little late to the starting gate and need helpful hints about how to integrate AI into your operations, rest assured that a watchful eye, a second thought, and a good gut instinct may lead you to draw assistive conclusions. A key thing to remember about artificial intelligence is its nascency.

 While it may be lightning fast and capable of surprising cleverness and accuracy, it’s still attempting to master its own imperfections. Generative AI is driven by what’s known as “machine learning.” Recent cases in point will quickly reassure any lingering doubts you may be entertaining about your own organization’s lack of technical prowess and may even prevent you from feeling left out or left behind.

                                 ADVERTISEMENT

Google up a quick image search of any AI-generated people and you’ll see some bizarre distortions of the human form. Even the nearly perfect likenesses often have deformed hands, with extra digits curling out of their knuckles. AI art-generating machines seem to be geniuses at constructing human faces but still sometimes struggle to accurately render certain lifelike bodily details. Think of artificial intelligence platforms as if they’re a small cohort of prodigy children who are light years ahead of their peers in some regards but still underdeveloped in other eventually perfectible areas of learning.

While artificial intelligence is good at analyzing frames of visual reference (mass data) from the World Wide Web and then regurgitating somewhat orderly images with attractive color, lighting, and detail, the user has to make a judgment call as to whether the output meets or beats their expectations and can actually communicate the desired message effectively to the audience one has in mind, whether for a social media meme targeting an influencer’s followers or a professional pitch deck for private equity investors. If not, one can always go back to the proverbial drawing board with a variety of more detailed verbal or written prompts.

Two things you can do to improve outcomes: either wait for AI’s imaging sophistication to improve over time or simply keep prompting the image generation machine with increasingly direct descriptions while weighing the images against needed results. This is the human reasoning and seasoning that AI lacks.

                                ADVERTISEMENT

The same goes for writing. Here are two sets of prompts that I keyed into ChatGPT to help generate ideas for this column:

Prompt #1

Write an 800-word essay about how AI is liberating executives in multiple disciplines so they can spend more time on core business functions, have more recreational time, and make a lot more money.

Prompt #2

Write a uniquely humorous 800-word opinion column about how artificial intelligence can liberate Guam-based business executives so they can spend more time on core business functions, make more money, and enjoy more recreation.

The first prompt generated several useful insights into how various business types and processes can be made abundantly more efficient.

But the column generated by the second prompt tried too hard to be funny with a series of corny, repetitive clichés. By the third reference to leisurely sipping coconut water on the beach while my AI machine made me millions, my wincing had turned to retching. To borrow a quote from early-2000s direct marketing guru Richard Harshaw, AI’s attempt at humor was like most predictable ad copy: “more stilted than the 11-foot clown at the circus.”

For now, go with what works. If your business systems are still personally fulfilling, rewarding for your employees and customers, and bottom-line profitable, simply start testing AI to find out which integrations can actually improve workflow, cash flow, and net gains. At this point, the modern world already knows that artificial intelligence is here to stay. And it will improve over time for all of us, the more we teach and train it through everyday use and experimentation.

Jeffrey Tomas Marchesseault is a broadcaster and real estate broker who now works in economic development, national and international affairs for the government of Guam. He loves to read, write, and riff about busted communication systems and how to fix them with technology. Send feedback to mediamaxguam@gmail.com

Subscribe to

our digital

monthly edition