Unveiling AI's Hidden Realities: A Journey into Its Shadows

While the buzz around artificial intelligence is thrilling, there’s a flip side that deserves our attention, according to Guy Courtin, VP of industry and global alliances at Tecsys.

This “flip side” refers to the staggering amounts of electricity and water AI systems consume. major players like google and Meta have even put thier sustainability initiatives on hold to push forward with AI advancements.

Take large language models (LLMs), as an example; they require an immense amount of energy just for training. Courtin points out that even a modest LLM uses enough power “to drive a car from Earth to the moon and back.” This raises an critically important question: shoudl we really be casually engaging with thes models? There are real implications here.

Moreover, AI isn’t without its flaws—hallucinations and inaccuracies frequently enough stem from poor data input. As Courtin puts it, “They’re just going to use it to give you an answer,” regardless of its validity.

AI can also become problematic when it provides accurate information that could be misused. For example, while ChatGPT won’t directly tell you where to find pirated music, if you ask which sites to avoid for such content, it might inadvertently guide you in the wrong direction.

The real magic happens when AI is used in specific contexts—like optimizing delivery routes within a warehouse where all variables are known. In such scenarios, users can expect more reliable answers since they understand their environment well.