What is the AI Revolution?

In my previous article – The Role of IT Architect in the AI Revolution – I tried to make the case as to why Architects will be critical to helping us successfully navigate the emerging AI Revolution. But if you ask 3 different people what today’s AI Revolution is or what it represents, you’ll likely get three different answers. Some of the more common answers might include:

1. The AI revolution is all about Generative AI and the big surprise that it was more competent and faster than anyone expected, or…
2. Someone might explain how we are on the cusp of actual AGI and that it’s an existential threat to humanity, or…
3. There might be a review of how Machine Learning has rapidly begun replacing tradition DevOps to provide capabilities that have been queued up, largely unfulfilled, since the era of large Enterprise Transformations began in the early 2000s. ?

Of course, we might consider that the AI Revolution is all of this and more. In fact, if we were to catalog all of the areas of potential use cases and impacts we’d quickly get lost in it and begin to wonder whether AI and IT were becoming synonymous. Well, they are or at least they will be. Perhaps a better way to quantify and qualify what this revolution represents is to tackle it in Capability Waves. When doing this, it’s important to recognize that we’re not in the first wave of AI Capability – far from it – but we are in the first wave of AI capability that’s passed a certain threshold (the Threshold that determines that we are in fact experiencing a revolution). Let’s take a look at this more closely.

What Constitutes a Revolutionary Threshold (e.g. one that defines entry into a revolutionary period)?

The criteria that can be applied to make that determination includes both generic and targeted considerations (targeted in this case, in relation to previously defined expectations associated with Artificial Intelligence). As we know, AI-related technologies have been around at least since the late 1980’s and the associated information theories much longer. Over the past ten years there has been a surge of AI adoption in particular areas, but that trend was more evenly paced.

2023 was different, it crossed the threshold boundary. That threshold included the following factors or criteria:

• The convergence of multiple Hype Cycles (in this case at least three, but probably more).
• Rapid adoption across a more or less universal array of domains (both business and architectural) of at least one and of the technologies associated with one or more of those Hype cycles – in this case Generative AI (although ML has also crossed most domains as well).
• Concerted Government and industry recognition and reaction on a global scale leading to immediate strategic and tactical redirection across nearly every functional domain.
• The surpassing of previous exploitation-related expectations, both in terms of capability available as well as the availability timeline. The realization of transformation impacts across a diverse spectrum of potential application areas (based upon the new expectations as described above).

Having said all of that, it’s important to recognize that we’re still in the initial wave – the one that begins just as and after this AI Revolution is acknowledged. We can expect many more AI Capability Waves in coming years and perhaps even another revolution (one directly associated with Artificial General Intelligence or AGI). It may be too early to define what the coming AI Waves will look like, but I think it may be worthwhile to try to classify them based both on their nature as well the resulting impacts. Other folks have tried doing this before – sometimes using a human age or an education-based analogy (grammar school, high school etc.) or using schemes associated with states of consciousness (aware, reactive, proactive) or multiple combined. The tricky part here is that several (seemingly different) things are likely to occur in each new wave and of course they’re all in the process of being merged together and merged into the larger ecosystem of IT capability.

Here is my initial guess at a classification scheme based upon “core generic competencies” and/or general outcomes for the AI Waves we’re likely to experience in the coming years.

Wave 1: The GenAI Wave or Augmentative AI (begun in 2023) – will last perhaps 3-5 years. While there are lot’s other things happening, the primary focus in this wave is going to be rapid exploitation of new GenAI-related technologies to accomplish tasks previously identified, but unfulfilled as well as a hyper-automation surge. As we’ve seen over the past year already, “Augmentative” can and does encompass creative endeavors. This does not make GenAI anywhere close to AGI, but it does highlight that hyper-automation can encompass most of what had previously been considered human-focused creative domains.

Wave 2: Autonomous AI - (likely to begin before Wave 1 is over) – will last 3-5 years. While some may argue that this could occur sooner, I tend to think that more advances are needed as well as a better regulatory and governance frameworks. Three to five years gives us plenty of time to make progress in both areas. It’s important to note here that unlike the previous wave, the range of capabilities (and associated AI technologies) will include a much wider cross section of tools and techniques – in other words, the spotlight on GenAI will subside somewhat and blend into a wider framework of related AI and emerging non-AI technologies (such as quantum computing, robotics, etc.). "Autonomous" in this context refers to handing more decision control to AI-based or focused systems once the proper governance and oversight (hopefully) has been applied (on a global scale). ?

Wave 3: Total IT Transformation - (also likely to begin before the previous Wave is over) – may last until we reach the next Revolutionary threshold. In truth, this is already beginning, but I believe it will take about 10 years to fully mature. In this Wave, all DevOps, all Cyber and most domain or mission specific applications become AI-driven as well as all productivity tools. As we get towards the end of this period, the distinctions between what is AI and what is IT will become meaningless, except in the context of AGI. In other words, non-AGI AI will drive all other IT capability, becoming both ubiquitous and transparent. ??

There is a fairly large bubble of uncertainty surrounding these timelines though, as some of the progress that’s occurring now seems to be exponential in nature and as we’ve seen recently – many of us we’re surprised that the first Wave occurred in 2023 as opposed to later – so the timeline represented here may be a bit conservative.

Whether there will be a next Wave or a new Revolution in relation to AGI is an open question and of course is fueling a lot of the existential panic referred to earlier.

Why is AGI different?

So, why would or should AGI represent a new revolution; why not include it as part of the AI Revolution? There are several reasons for potentially separating them:

1. It is difficult if not impossible to predict at this point exactly what capabilities may emerge from AGI. Just as GenAI surprised many, AGI will likely surprise everyone (and of course that’s why so many people are sounding the alarms now before it’s upon us).
2. Introduction of widespread AGI will likely have a huge impact on society; on the economy, on politics, on government policy and regulation – on basically everything. Entirely new strategies, standards and frameworks will be required to deal with it – those will not emerge for some time yet. ??
3. AGI will likely be viewed differently from all previous technologies. In other words, once AGI has truly been achieved, the differences between humans and AI entities will become difficult to determine; especially when it begins merging into robotic and genetically engineering living platforms (e.g. anthropomorphic forms). This has always been the stuff of Science Fiction until now, but fiction and science fact are on a collision course for merger in this case. Once our technology begins to demonstrate human characteristics, the question of rights must be considered and the fundamental relationship between us and our technology will change.

As noted in the previous article on IT Architecture, all of this evolution will require some serious, holistic analysis – which means that people who have experience assessing complex, dynamic environments need to be in the forefront in order to successfully navigate all this. There is a more risk and more opportunity associated with the AI Revolution/s than our previous IT revolution and likely as much as the Industrial Revolution itself. In relation to the Industrial Revolution - we have came close to realizing several existential threats – everything from Nuclear energy to Global Warming. The AI revolution inherits all of those threats and bundles in several new ones.

It’s going to be an interesting next 10 years or so. Much more interesting than the past decade.

?

Copyright 2024, Stephen Lahanas