The Sigmoid Curve and AI: where are we?

AI is swiftly becoming one of the most transformative technologies of our time, reshaping industries, enhancing productivity, and having the potential to revolutionise the way we live and work. As we navigate this ever-evolving landscape, many of us are wondering what’s next? This is where the sigmoid curve can be applied to measure where we could be in the current state of AI development. The sigmoid curve, also known as the S-curve, is a mathematical representation of the life cycle of various phenomena, including technology. It illustrates the progression from the initial stages of slow growth to rapid expansion, followed by a plateau and an eventual decline or transformation phase. When applied to AI, this concept can help us to gauge where we stand in terms of AI development, adoption, and overall progression.

The initial stages

In the early stages of AI, progress was relatively slow. Researchers were exploring the fundamental principles and developing the building blocks of AI systems. The first ‘AI winter’ occurred in the 1970s which marked a period of reduced funding and a general sense of disillusionment, as the initial hype surrounding AI did not live up to the expectations put on it. However, breakthroughs finally came in the 1980s and 1990s which reignited interest in the technology and set the stage for the rapid growth phase of AI.

Rapid growth The late 1990s and early 2000s witnessed a massive surge in not just AI development, but applications for said developments. Particularly, areas such as natural language processing, computer vision, and machine learning took off during this time period. The exponential growth of data, coupled with advances in general computational power fuelled this expansion. This was also the first time period where we saw companies adopt AI technologies for tasks such as speech recognition, recommendation systems, and fraud detection.

As we move into the 2010s and 2020s, this is where AI has really begun to take off, with a seemingly constant stream of new advancements coming to the industry. This phase of growth, along with that seen in the late 90s and early 00s is often referred to as the ‘golden period’ on the sigmoid curve.

Is this only the beginning, or the start of the plateau?

Now that we find ourselves on the cusp of a whole new phase of the AI journey, where are we on the curve? We are now witnessing the challenges and limiting hurdles of existing AI technologies, such as biases in algorithms, a lack of interpretability, and ethical concerns – so is this the plateau, or is this still just the beginning? Whilst these issues are certainly leading many of the industry’s greatest minds to question the development of AI and raise a focus on responsible AI development and deployment, it is certainly not the end of the cycle just yet.

Instead, it could be that we’re still only at the beginning of the curve, with yet to relish the full-scale potential of AI, or more likely this plateau is simply a ‘transformational’ one, and more growth is set to follow soon enough.

The plateau phase presents a unique opportunity for reflection and refinement, which is what we are currently witnessing in the AI landscape. Researchers and practitioners are actively working to address the shortcomings of AI systems and make them more robust, transparent, and trustworthy. Interdisciplinary collaborations are also continuing to emerge, bringing together experts from diverse fields to tackle complex AI challenges. The emphasis is shifting from just building AI systems to building AI systems that are more aligned with ‘human’ values and societal norms\needs.

This phase can often present the previously mentioned potential for transformation. As AI matures as a technology, it has the potential to redefine entire industries, create new business models, an reshape the job market. This transformative phase is often characterised, when the sigmoid curve is applied elsewhere, by disruptive innovations that challenge existing norms and create new opportunities. This is something that we are continuing to see the impact of in sectors such as healthcare, finance, transportation, logistics, writing, art, and manufacturing.

However, it is important to strike a balance between the hype and reality of AI capabilities, for both the positive and negative sides of the argument. Whilst AI seems to hold tremendous potential, it is not a ‘panacea for all woes.’ The sigmoid curve acts as a reminder for us to be cautious and uncertain of potential, whether that be in regard to inflated expectations or future dangers – lest we slip into another ‘AI winter.’ Only by doing so can we ensure a future where the benefits of AI are not just accessible to all, but also come with minimal risk and unintended consequences.

As a final note, the sigmoid curve acts as a valuable framework to assess the current state of AI development. We have now moved past the initial stages of slow growth and are witnessing the period of rapid growth which follows – the only question being where we are on that curve. As we navigate forward, we must now focus on responsible development whilst also addressing the challenges that come with widespread AI deployment and advancement. Only by doing so can we pave the way for the transformative potential of AI to be realised whilst ensuring that it serves as a force for good in our evolving world.