I gotten from a Mandarin Chinese YouTube video on Artificial Intelligence.
"AI产生智慧并不可怕,可怕的是因为AI的出现,人类从此丧失了思考和持续学习的能力。”
"It's not scary that AI generates intelligence; what's scary is that with the emergence of AI, humans lose the ability to think critically and continue learning."
While it is scary, it is also very plausible as well.
Think about it, since our hunter-gatherer days, only body that manage to conserve the most energy together with efficient use of energy will survive in an era where food are limited and everyday is constantly a trade-off between expending consumed food to hunt for more food for survival.
This survival process programmed humans to move less where possible and thus humans do not really enjoy exercising. When was the last time you exercised, not out of habit? (If you are interested in reading about this topic, I recommend “Exercised” by Daniel Lieberman)
What does this mean? Two things:
We humans has a strong tendency to be lazy.
Brains are like muscles, they consumed energy too while thinking.
While we are at the stage where AI has proliferated into business processes, chances are high that humans-in-the-loop is still needed in most of them. Unless the business can trust that the AI built is aligned with the business, which might not be the case, especially connectionist models. For instance, in an augmented decision making scenario, where the machine provides recommendation action and humans makes the final say.
Let’s look at ChatGPT. Many will have remembered the earlier days of ChatGPT, where it hallucinates a lot, especially with reference materials. It got a group of lawyers into trouble as they did not double check the reference made by ChatGPT and present it as it is.
Imagine you are a judge, and a COMPAS-like recommendation engine is provided. Although it is sold to you as a decision-assistance tool, but given that you are inundated with so many cases especially if they are pretty trivial, there is going to be a strong temptation to solely rely on the recommendation engine’s recommended remedial and mete out the sentence accordingly.
My point is, to all product managers, innovation managers, business managers, please remember humans can be counted on to be lazy, complacent and very dependent. This is something that will not change in a very long time. So when you design any process where you are putting AI together with humans, please always remember that and design it accordingly.
What are your thoughts on this?
Do you know of any product managers or business managers who are building AI products? Perhaps forward this newsletter issue to them to raise awareness!
Please support this newsletter by liking, and forward it to friends who might be interested! Thank you! :)
This is the first legitimate fear I have since the fear-mongering of AI started almost immediately after ChatGPT. An average human will turn into ones we see in Wall-E and that is a real, existential threat as a species.