In the history of the development of artificial intelligence technologies, periods of AI winters are periods where the domain of artificial intelligence saw reduced funding and reduced interest. This includes both the interest in products that leverage these technologies as well as the research in AI technologies.

In the first part of this article, we started covering the drivers behind the first two AI winters. This part continues that discussion.

Such was the stigma associated with the term “AI” during the second AI winter that researchers in the mid 2000s deliberately tried to name their algorithms in ways so that those algorithms would not get associated with artificial intelligence. Machine learning was a term that was a result of this quest. Some of the other terminologies used were informatics, knowledge based systems, business rules management, cognitive systems, intelligence systems, intelligent agents or computational intelligence.

Since 2015, the world has largely seen a period of AI spring. What this means is that we have seen a rapid interest in AI related technologies emerging again. While many may not be able to make a direct link, the fact is the advances in AI that we see today somewhere originated because of the launch of Apple’s iPhone. Among many other things, what Apple’s iPhone or Steve Jobs’ vision demonstrated is that what was considered or viewed as futuristic till that point, could actually be brought into vision with technology. 

Technology could actually act as a bridge between idea and reality. The fact is, Jobs demonstrated to the world that if something is possible theoretically, and technical capabilities theoretically exist to support the vision, with the help of some of the best minds in the world, that vision can be turned into reality. This development actually encouraged many other entrepreneurs as well as researchers in academia to be more bold and innovative about their ideas. 

In my perspective the current AI spring, which may be now transitioning into AI summer, is largely a result of the introduction of the first iPhone. In fact many of the technologies that we saw get integrated in Apple’s iPhone progressively, were actually within the realms of technologies that were considered to be a part of the artificial intelligence domain. But the fact is that if you analyze closely, you will find that since 2022 the pace in terms of innovation has slowed down a bit. You may think that this is the most inaccurate statement ever. So allow me to explain what I mean by this. The innovation aspect pertains to a new capability (like LLMs) emerging. Subsequent modifications in capability, size and modalities are technological progress, not innovation.

These technologies will obviously become more powerful and more capable, and get integrated with other emerging technologies but in terms of something new cropping up, we may see a lull for next couple of years. And that is the reason I say that we are entering into AI summer. During the spring, with these new and exciting technologies coming on the horizon, there has been so much hype and proliferation of products and companies in this arena that the level of hype is unlike that we have seen before. AI summer indicates the need for a phase of maturity where the hype needs to turn into reality.

One thing that has been significantly different, when compared to eras of previous AI winters, is that this time the hardware technology that can support building the capabilities exists. And these technologies will only mature to become more and more powerful. Ideally, the true value of an AI product would be to create a capability that did not exist before. Let us use the example of natural language processing (NLP). The fact is natural language processing already existed before generative AI came into picture. You could do things like translation summarization using other approaches bucketed under machine learning.

So if a generative AI tool comes and makes these capabilities much more accurate, or more efficient, it’s not exactly innovation. It is a transformation from the current state. However, when it comes to generating videos and pictures through text, this is a capability that did not exist before. Neither did the capability to ask an algorithm to perform some sort of analysis, whether theoretical or numerical, through dialogues. But these capabilities which are new are still at their “demonstrated reality” phase.

What this means is that while it has been demonstrated that a plethora of things can be done and can be converted into a productionized solution,actual commercial uses have been limited.On one hand we have big tech companies that have these solutions that can integrate or have already integrated these solutions or these advances into their existing portfolio.

Again, while this significantly enhances their existing products, it has not led to an innovative new product line altogether. On the other hand companies that focus just on these technologies or have focused on just these technologies, have not been able to successfully take this from the initial demonstrated reality, beyond AI enabled automation.