The AI Revolution Is Not Here Yet

Unrealistic Expectations from AI

This is not the article trying to be the voice that AI won't matter, that it won't impact businesses or eliminate jobs, or that it is not already having an immediate impact on productivity because it already is.? AI is already making workers 10-15% more productive. AI will impact how we live, work, and play (being a gamer, the author of this article is particularly excited about what AI can do for virtual worlds).?

However, much akin to the computer revolution of the 80s and Web 2.0 that wasn't and then was, AI is way, way overhyped. FOMO (fear of missing out) is rampant, and people who make decisions based on fear rarely make good ones. Much akin to the technology mentioned earlier revolutions, which both promised (and I was there, so I am not making this up) to eliminate poverty, war, and wealth inequality, bring down all oppressive governments, eliminate all diseases, and stop any future pandemics and bring about a supercharged scientific period akin to the Enlightenment that would achieve the lofty goals of eventual immortality and overcoming gravity and space-time itself.??

Well, that didn't work out so well.?

If this sounds familiar, it is because the proponents of AI are making the same promises. Humans have this inability to learn from the past. The cycle of technology since the 60s has been creation, hype, adaptation, reality, disappointment, deployment, and effectiveness. AI may cycle through this faster than any previous technology advancement. But it will go through the cycle. When AI reaches the effectiveness stage (where it is ubiquitously adopted by society), it will not look as we predicted. I can say this with confidence. Why? Because none of the current effective technologies were integrated into society as anticipated.

Worse, the unanticipated usages and human adaptations of technology cause the most problems (technology is and will always be about humans). Social Media has benefits, but the costs have been high, causing depression, anxiety, fear, anger, and tribalism among those most addicted. It has also been used to silence opponents and spread false or misleading information. And wait till AI gets a hold of SM.?

AI in its current form has many limitations that will limit it from reaching the practical stage anytime soon. AI has issues with hallucinations. Are we going to allow AI to run accounting and be a lawyer or doctor when it tends to make things up when it does not know the answer (which, ironically, is human-like behavior)? A lawyer filed a brief in a lawsuit using ChatGTP. It did not go well. Radiologists and doctors are catching medical AIs making fatal mistakes. While AI working with doctors will improve health care, we have yet to determine if we will ever get to the point where doctors are unnecessary. My prediction is that we will not.?

Another example is coding. Can AI replace developers? Not really. It is making them more efficient, but there are many complex problems it cannot solve. A friend recently texted that he used Microsoft's CoPilot to create a simple game. I responded by asking him to make another RimWorld. AI still struggles to solve complex coding problems.?

AI has other limitations. AI lacks context. AI is not contextual, and this limits its abilities. Humans do this without thinking, so even the more sophisticated chatbots currently using Bard or ChatGPT can quickly reach the end of the conversational rope. They are much improved over the canned versions, but if you have interacted with one for an extended time, you will eventually get transferred to a human agent.

AI faces challenges in gaining user and organizational trust. As a recent Harvard study showed, implementing?AI in the workforce can be a dual-edged sword.?

If so, how do you integrate AI into your organization? How can you reap the benefits? I will continue exploring those excellent questions?as we integrate AI into existing companies.?

Stay tuned.