AI is more than just a hype

AI has been the buzzword of the day. The number of organizations and professionals aiming to be AI-Native, or AI-Savvy or AI-choose-a-fancy-word is reaching new heights. The tiny country of Angulla whose national domain name is .ai is boosting its revenues as many companies rush buy .ai domains.

Yet, experienced professionals and organizations have seen these hypes come and go. Don't get me wrong: yes, AI is is fantastic tool that may be a great generator of competitive advantages if properly applied in the right context to support business objectives.

There is life beyond AI

However, we shouldn't believe IA is "one size fits all".

First, because AI is an umbrella term for a set of different techniques. Second, because not all problems can or should be solved by using AI. There is the case of a manager who asked his IT intern to develop an AI solution to detect and replace all the domain names in data flow. To his surprise, after two hours, the intern came back to him saying "it's done". The manager asked "What do you mean it's done?" as he expected this to be a lengthy task. The intern just showed him his solution, which was based on regular expressions. The manager was surprised by the simplicity and elegance of the solution. For the IT intern, this task was just a case where using AI would be just an overkill solution, and a simpler, yet very effective alternative solution could be used.

This story highlights three important facts:

• Selecting a predefined approach without considering a wider set of solutions is not wise
• AI may not be the most economical or effective mean to achieve a specific objective
• You need to have the adequate background, knowledge and experience to select the best approach and distinguish between the hype and the best solution
• The difference between a wise choice or a follow-the-hype choice can be measured in time and money

AI is just another (very valuable) tool amongst a lot of other good solutions. Ultimately, the business problem should be the primary driver of the selection of the right tool.

AI is not something new

Many AI techniques we use today were already developed in the early 1960s. There was also the famous case of the Japanese project called "Fifth Generation Computer Systems" [1], which was motivated by the fact that (at the time) "Current computers are extremely weak in basic functions for processing speech, text,graphics, picture, images, and other non-numeric data, and for artificial intelligence type processing such as inference, association, and learning". Ultimately, the project was not highly successful as the computing power available at the time was limited, and the program didn't make extensive use of core concepts such as the Neural Networks that are the core of todays LLMs such as ChatGPT, Copilot, etc.

In fact, in my MsC in Computer Engineering the 2000's, we were learning these AI techniques (ex: neural networks, classification models, etc) together with a wide set of other equally relevant approaches - such as metaheuristics, automata theory (where the regular expressions used by the aforementioned IT intern), etc), decision trees, etc). It was already apparent to me at the time that there was no best approach - each one had their optimal field of use. This perception was reinforced throughout my career both in my PhD research and in consulting work, where i had the chance of applying many of these approaches and understand the right contexts for each of them.

The popularization of LLMs such as ChatGPT and their apparent ability to "reason" led many people believe that AI was a brand-new invention and the answer to every problem. What happened is that AI jumped from research labs to the lives of everyone through LLMs - it has gone mainstream. We evolved from the (now ancient) mantra "Google knows it" to "AI knows it". But a pragmatic analysis of the problem is always a better approach than following a mantra. 

How we can help you

At Mourinho Solutions, we follow no mantras. We have the knowledge and experience to help you choose the right approach and the most effective tools to ensure your business thrives. We analyze your business needs to pinpoint where AI (or other approaches) can deliver the greatest impact, giving you a competitive edge.

Your success is our success too. Don't hesitate to contact us to know what we can do for you!

References

[1] - Feigenbaum, E., Shrobe, H.: The Japanese National Fifth Generation Project: Introduction, survey, and Evaluation. Future Generation Computer Systems. 9, 105–117 (1993). Available here.