TECH TALK
Shall We Play a Game?
Diving Into AI
By Tak Sato
When my wife and I started dating in college, one of our favorite pastimes was watching movies. The weekend meant a pilgrimage to the theater.
After our son arrived, our weekend routine began at the Blockbuster store to rent new releases. And now? I pop the popcorn while she starts to stream the movie. Thanks to artificial intelligence (AI), Netflix’s AI learned our taste in movies; its recommendations are spot on.
Getting Real
“Shall we play a game?” is an iconic line by an AI in the movie “WarGames” after a young hacker named David Lightman (Matthew Broderick) unknowingly breaks into a supercomputer at North American Aerospace Defense Command (NORAD) that simulates nuclear war at the height of the Cold War.
Total annihilation of both sides is the result of the nuclear war simulations (games) until the AI “learns” that not starting a nuclear war in the first place means victory.
Artificial intelligence, one of the many disciplines within the study of computer science, has been around since the ‘50s, mostly behind the scenes.
For example, credit card companies use AI recordkeeping to analyze billions of credit card transactions and identify patterns that are likely fraudulent. It’s one way AI efficiently and accurately outsmarts a customer service representative’s intuition (organic intelligence).
Here’s another example: A traditional Clinical Decision Support System (CDSS) analyzes a patient’s medical records to provide treatment options so a physician can recommend the best treatment.
During a panel discussion entitled “AI: Friend or Foe” at our annual Living in the Digital World Senior Expo & Conference last year, we learned that CDSS is incorporating and strengthening AI. With AI baked into their computational processes, CDSS can cross-reference, for example, a cancer patient’s CAT scan against thousands of other CAT scans with similar cancer staging, therapies used, prognoses, and actual outcomes.
CDSS supports physicians’ critical thinking and ultimate decision-making for treatment, something that AI is not capable of doing; at least not yet.
Until the fall of 2022, AI primarily benefited businesses directly and consumers indirectly. ChatGPT is a product of an American AI laboratory OpenAI, and it ignited the AI boom, bringing AI tools and services into consumers’ devices.
Tech titans like Google and Microsoft entered the race with a vengeance and urgency. As I pen this article, a Chinese company released DeepSeek with its initial claim to fame—that it was a less expensive AI system than American versions such as ChatGPT from OpenAI, Google’s Gemini or Microsoft’s CoPilot.
ChatGPT falls into the AI categorization, known as generative AI, that responds to commands or requests. If you type in “Write a poem about our Rhodesian Ridgeback and boxer puppy,” it will create a grammar-perfect poem with those references. Some regenerative AI systems use images and videos instead of text. Entering the prompt, “Show an image of a flying pig with bright red lips and a long snout carrying a cowboy” will quickly give you that image on your computer.
But Wait…
When you Google something, you’ve likely noticed an AI-generated answer at the top of the website links. If you are comfortable with the AI-generated summary, you may skip the website links to find your answer. As a geek who coaches people on safe digital habits, an AI summary can keep you from accidentally going to a bad link that opens a malware-infested website.
I’ve just scratched the surface of AI. Let me conclude with something my son learned at his middle school’s National History Day competition. Although AI has the advantage of learning and analyzing based on what it’s “learned,” that information may not be reliable.
My son was told to check his primary sources… good advice for everyone using AI.
Consider this: if AI could determine truth from fake news 100% of the time, why would Google lawyers put “Generative AI is experimental” at the end of its AI-generated summary section? I gotta believe in a lawyer who makes tons more money than me.
