I hope you enjoyed the “first period” of this blog article. If you have not read it, you can click here.
Crucial Un-Artificial Intelligence
Artificial intelligence is built from its own cognitive learning process coupled with massive storage and computing power to complete an analysis much faster than humans. However, can AI think better than humans? With the inductive (vs. deductive) reasoning as intelligence gathered over time. Can AI integrate or replicate the human element?
Where does AI gain the hockey “knowledge”? The human equation?
My former students will (may ha-ha) remember this expression from my classes …
Technology systems can only be as effective and operationally accurate as the experienced professionals who design and implement them.
Like coaches and athletes, their brain is a human AI for sports. Earned intelligence. My definition is … a body of knowledge, experience and neural processes (genetic and learned) that provide the capabilities to analyze and make informed decisions.
So, does technology move from a process to implementation directly without human intervention? For strategic and tactical activities … the simple answer is no. For some operational activities … yes or possibly. However, the process must be validated. Against baselines as well as the current issue. Question … Would you want your next airplane flight to be 100% operated by AI without a pilot in the cockpit? Without flight attendants?
Respectfully, I would suggest no. It appears that leadership (corporate, health care and education) are totally laser focused and “dialed in” on AI. For the quick answer. To reduce costs. While increasing customer satisfaction, product design and production and providing services. I would respectfully recommend stepping back and guiding the implementation in a thoughtful and deliberate manner. A question to ponder …
Have you ever been frustrated and disappointed by the infusion of technology as a customer? As an employee while performing your day-to-day activities?
Domain Knowledge
A few decades of knowledge (combination of "real world" and academic) in technology development and implementation have taught me one undeniable fact.
Domain experts must properly design, develop, test, implement and monitor to introduce an effective information technology system.
Notice that I referred to domain experts and not technology experts. As an example. Designing a car requires experienced and skilled production personnel. However, it also requires the same in other functions by gathering and analyzing data of a product:
- Marketing: Customer, survey, industry, and competitive data to gather trends and interests.
- Finance: Financial, cost, and industry data to assess the quantitative values.
- Strategic: Long-term trends associated with the organization’s future and value proposition.
- Engineering: Production, technical, and operational feedback data.
- Supply Chain: Cost, availability of raw/finished goods/sourcing, and production feedback.
OK, it’s summer … now let’s try a deep dive into the System Model.
Feedback Loop
The typical system model, taught in many business core courses, was defined decades ago. The system model is not necessarily defined as a technological system, but as a system of processes, often translated into a technology system. For a definition and example of this mode click here.
In a traditional organization framework: 1) How is this information gathered? 2) When? 3) From whom? The feedback loop “transmits” the output from any process becomes the input to the process (or other systems). AI experts will voice that this process is the premise of generative AI. Of course!
In Part 1 of the blog article, it discussed the integration of AI into a hockey club. The feedback loop would consist of the following:
- Input: The game strategy provided by the AI model.
- Process: Playing and completing the game.
- Output: Result (win vs. loss) as well as video and statistical data from the game.
Did I leave anything out? Yes, the feedback loop. Is that simply the data from the output phase directly “feeding” the input without any human interaction? If so, there is no human input. Maybe, the competitive off-line data for an opponent’s player will not play (injury, trade, retirement).
If organizations rely on AI to be the process, then there is no human intervention. The only human intervention are players playing the game, otherwise completely automated. That would reduce the club payroll, wouldn’t it?!
Circling Back to the Importance of Process
If you read to this point, you may be saying … “Doc (if you are a former student or teens at my parish) is forming a case to say that the process (game management) cannot be fully automated!” Not really, there is another point to this discussion.
OK, the third period will start after an intermission. Go to the concession stand to get some refreshment. And a revelation that may shock some people (especially former students!). 😊
If you wish to print a copy of this article, AI--OpportunitiesAndChallenges_2024-Aug Part2.