Many of us may not be aware that the term “machine learning” came into existence to avoid the stigma associated with the term “artificial intelligence” during the second AI Winter. It may be difficult to fathom in today’s age of AI hype that AI once was a word that many business investors and those funding academia, did not want to hear. So many researchers, who started their research within the realm of artificial intelligence decided to re-state the topic by coining other terms like computational intelligence, machine learning, cognitive intelligence, etc.

When machine learning regained prominence, and the AI spring began in 2015 after the second AI winter, we conveniently began defining machine learning as a gamut of algorithms that fall within the realm of artificial intelligence. While the most recent technologies in the AI realm, such as large language models, certainly provide possibilities that have not been explored yet, the current hype around artificial intelligence has been propelled and provided credibility solely by the success of machine learning algorithms.

Since the late 2000s, machine learning algorithms have delivered immense value in both academia and industry. The use of these algorithms in transforming analytics and automation approaches in the industry is what established the credibility of machine learning and subsequently, the term AI. When machine learning delivered value, we emerged from behind the shield of machine learning and proposed the term “artificial intelligence” again, while obviously expanding the realm of what constitutes artificial intelligence. But since this time we have been able to deliver value through machine learning, the world, at least at this point, has happily accepted the term “artificial intelligence” and believes in the opportunities it promises.Keep in mind that the umbrella of machine learning spans a wide gamut of algorithms, including the widely familiar algorithms like regression analysis.

That is why we have actually not even explored the full potential of what will be considered classic machine learning algorithms. Survey data from many research firms estimates that a majority of organizations have not actually successfully leveraged the true value of even the full gamut of machine learning algorithms. But the gist is that the current popularity that the term artificial intelligence enjoys is due to successes in machine learning.

And that is the reason we have to be careful with the current AI hype. As we push the intelligence aspect in the term artificial intelligence to new heights, we risk navigating into many different scenarios, none of which are good. We keep emphasizing that the progression of artificial intelligence is moving toward an objective of a very vague term called superintelligence.

But no one is sure what defines superintelligence. My article Mitigating The AI Effect captures this dilemma or paradox, whatever you want to call it. The reality of this path is that after a while, fatigue will set in where people, the masses, and the investors will not see what will be their definition of superintelligence or artificial general intelligence (AGI). If the imagination of superintelligence is the kind of machine we have seen in science fiction movies, that is definitely not going to happen. The gist is that we are propelling the hype toward concepts that are not yet concretely defined.

The other dangerous scenario is that all this attention on large language models might shift attention away from other algorithms that have until now been defined as machine learning algorithms. Remember that 80% of the value from these machine learning algorithms has not been extracted yet. So there is this massive opportunity that still exists in the realm of machine learning algorithms. And then on top of it, these machine learning algorithms can be integrated with what is currently being termed true artificial intelligence, like large language models, for more innovative solutions.

So what is the way out? How should we treat the path? We will cover this in the upcoming Designed Analytics Report, which aims to avoid a third AI winter. The report will be published on July 8. Stay tuned!