GenAI or something else?

GenAI or something else?

In recent conversations with industry peers & colleagues, a recurring theme has emerged: the intersection of environmental considerations and Artificial Intelligence. 

The rapid advancements in AI, particularly in Generative AI models, have opened up a plethora of opportunities. These models can consume vast amounts of data, such as entire documents, and extract very specific information like the parties involved. But the pressing question is, just because they can, should they?

There have been tech stacks capable of understanding documents and using Natural Language Processing (NLP) to extract relevant details for as long as I can remember. 

These technologies, while perhaps not as "advanced" as Generative AI, have been efficient and effective in their tasks. The introduction of Generative AI models to perform tasks that existing technologies can handle raises concerns about the environmental footprint of our choices.

Generative AI models, while powerful, are notoriously energy-intensive. Using them for relatively minor tasks, like extracting parties from a document, might be seen as overkill, especially when considering the environmental implications. The carbon footprint of training and running these models is significant, and in the context of global efforts to combat climate change, every decision counts.

While Generative AI can be invaluable for prototyping and testing Minimum Viable Products (MVPs), it's essential to recognise their environmental cost. As innovators and technologists, we must consider the long-term implications of our choices. 

There's a very real possibility that future generations (and many right now) will look back on our indiscriminate use of power-hungry AI for trivial tasks with dismay. The conversation around environmental impact and AI is a timely reminder of our responsibility. It's not just about what technology can do, but what it should do. 

As the AI landscape continues to evolve, it's imperative to prioritise sustainability and seek out less energy-intensive alternatives for the long haul.