The Rise of Artificial Intelligence

By 2030, artificial intelligence (AI) is predicted to contribute nearly $15.7 trillion US dollars to the global economy. That’s roughly the combined GDPs of Germany, Japan, India, and the UK.

AI is so widespread that in November of 2022, ChatGPT reached 100 million active users two months from launch – a feat that took Instagram 2.5 years.

This surge in money, time, and labor associated with artificial intelligence, coupled with the apocalyptic-style conversations around it, has been overwhelming and stark to many.

In reality, generative AI programs have been adopted at much wider rates than any past technology. PC’s (personal computers) and the internet, had a 20% adoption rate over the course of 2-3 years. But AI’s adoption rate is 39.5% over that same period.

Part of the increased acceptance around AI programs is that we don’t always know when we’re using them. Generative AI has been integrated into phone call systems, customer service bots, and even google searches – so oftentimes, we may be using AI without knowing it.

Additionally, AI has been incorporated into essential programs for both school and work.

Microsoft has created an AI system dubbed ‘Copilot’ which has the capability to analyze data, take notes, and edit for PC users worldwide. And every search a student (or worker) makes on Google generates an ‘AI overview’ – which is often used word for word in written (and oral) explanations. Usually, those features are not ones you can omit from your viewing, but rather ones you make an active choice not to utilize. 

The biggest conversation around AI for many is the issue of plagiarism and labor replacement: AI is stealing from online databases in an effort to replicate human work to become just as good as those original authors/artists/researchers, and that stolen work is used by other humans to cheat the system and make their lives easier. 

While those issues are significant to our futures, some larger problems stem from the energy usage behind Artificial Intelligence programs. Not only does the labor behind AI design cost hundreds of thousands of dollars (if not more) – but the training behind its task is detrimental to the environment.To ‘train’ OpenAIs GPT-3 took “1,287 megawatt hours of electricity (enough to power about 120 average U.S. homes for a year), generating about 552 tons of carbon dioxide.” And once AI is trained, each query (for OpenAIs ChatGPT) “consumes about five times more electricity than a simple web search.”

Each web search takes up about 0.2 grams of carbon dioxide. Which doesn’t seem like very much – until you realize that would be 1 gram for each ChatGPT query. ChatGPT processes more than 1 billion queries every day – for a grand total of just over 1102 tons of carbon dioxide emitted by ChatGPT every day. And that isn’t even considering the amount of water, physical resources, and time put into building the largest data centers the world has ever seen, the ones whose outputs continue to exponentially jump annually, for the storage of AI softwares.

It also doesn’t take into account the various other AI servers, chatbots, and programs out there.

So understanding the ‘risks’ of AI is important, and notably necessary to understanding the intelligence behind what humans do daily, but sometimes being so focused on WHAT artificial intelligence can makes us oblivious to where it comes from. 

And that will eventually, if not already, take large tolls on our environment. 

Reducing personal AI usage – whether that means not using ‘AI Mode features’ or chatbots in the place of a web search – is a small step forward in reducing the total energy expended by AI systems worldwide. For businesses, or larger corporations, continuing to use and pay for human labor (instead of an AI program), is key. Over time, the less AI usage there is, the less resources will be expended to its development. So eradication of both issues starts with a stop in AI usage overall. 

And every query avoided counts in the journey to lower AI’s ravenous consumption of natural resources and human opportunities.