Adaptive AI: the new technology that adapts to the environment

Artificial intelligence (AI) surprised the world during 2022: it ceased to be a laboratory concept and leaked into everyday use with impressive applications such as Dall-E, capable of generating images from scratch based on users’ natural language commands, or the more recent ChatGTP, with its infinite capacity to write texts, follow conversations and even write songs. Among the technologies ranked as key trends for 2023 is a new evolution: adaptive AI.

Unlike traditional systems, and as its name suggests, adaptive AI adapts to changes in the environment by reviewing and modifying its own code based on the appearance of elements in the real world that had not been previously contemplated during its programming. Thus, it could position itself as an ideal ally for companies that need to be adaptable, flexible and resilient in these times of uncertainty and disruptions.

Gartner, the market consulting firm, predicts that by 2026, organizations that have been able to adopt AI engineering practices to create and manage adaptive AI systems will overcome those that have not by at least 25%.

Unlimited use cases

To work, adaptive AI combines agent-based design and reinforcement learning concepts to “learn” patterns of behavior based on previous human and machine experience. It is based on principles such as robustness (high accuracy in algorithm results), efficiency (low resource usage), and agility (dynamic change of its operating conditions based on changing requirements).

Applied to customer care, adaptive AI could improve with each interaction its conversational skills, the understanding of customer needs and the understanding of the environment to deliver responses with unprecedented levels of accuracy, driving engagement and satisfaction to new levels. Another use case can be with demand forecasting: in the face of a severe disruption, it could quickly rethink its estimations and make the appropriate adjustments.

It also offers remarkable results in the dynamic adjustment of computing needs in data centers and edge computing structures: it distributes edge application workloads precisely according to the environment and requirements at any given time, generating significant savings in energy and resources by processing capacity and memory.

There seems to be no limit to the number of possible applications. Gartner cites the example of the U.S. Air Force: a learning system that adapts its lessons to the learner’s individual strengths. It thus knows what to teach them, when to assess them and how to measure their progress, as if it was a real individual tutor. This takes the levels of personalization to the maximum.

Philosophical challenges

From the point of view of its incorporation in companies, adaptive AI states profound philosophical challenges, ranging from a rethinking of decision-making models to a detailed evaluation of the ethical implications of its use. Regarding the first point, to deliver maximum added value, adaptive AI should be handled with a reasonable level of autonomy. Companies will therefore need to find a balance between their traditional decision-making models and the disruption that this technology could generate. For all these reasons, it is essential to reinforce change management in the face of an adaptive AI implementation: if there is no real change in the organizational mindset, it will be difficult for it to succeed.

Artificial intelligence is, by its very nature, surprising. And everything seems to indicate, with adaptive AI at the forefront, that the best is yet to come.