Day in and day out, we continue to hear about artificial intelligence (AI). Whether it’s a new company introducing a faster, more cost-effective model or an AI capable of generating hyper-realistic images that blur the line between human and machine, AI is advancing at an astonishing rate.
But what exactly is AI, and what impact might it have on you? Further, what is the impact of AI on our planet? These are the questions I will explore as we examine AI and its far-reaching consequences.
WHAT EXACTLY IS AI?
To gain insight into this question, I thought it would be interesting to see how AI defines itself. When I asked ChatGPT, “What actually is AI, and can you explain how it works in simple terms?” This was the response:
“AI is the ability of computers or machines to perform tasks that usually require human intelligence. This includes things like recognizing speech, understanding language, making decisions, and learning from experience.”
ChatGPT drilled deeper into how AI works. To summarize and paraphrase the output, it told me that AI is at the heart of many innovations, but it all begins with data—and a lot of it. AI learns from diverse inputs like images, text, numbers, and more. Using this data, it processes information through algorithms to uncover meaningful patterns.
Over time, AI refines its approach by learning from mistakes and continuously improving. This allows it to make smarter decisions, like recommending your next favorite song or recognizing a face in a photo. From simple chatbots to cutting-edge technologies like autonomous vehicles, AI applications are diverse and impactful.
So what is behind AI? The answer to that is algorithms, specifically:
- Rule-based algorithms are the simplest, following predefined rules like “If driving, buckle up.”
- Machine learning (or ML) algorithms, which are more dynamic as they learn and improve on their own without constant reprogramming. Forms of machine learning algorithms include supervised learning, where the system learns from labeled data (for example, spam classification); unsupervised learning, which identifies patterns in unstructured data (for example categorizing shoppers by habits); and reinforcement learning, which improves through trial and error (for example, chatbots learning from user feedback).
- Then there’s deep learning, which uses artificial neural networks to mimic human brain processes, enabling AI to tackle complex tasks like powering voice assistants or self-driving cars.
Though its mechanisms may seem complex, AI is fundamentally about learning, adapting, and improving through data and algorithms.
While this is all interesting, there are hidden costs and implications of AI to our livelihoods and the planet.
WHAT’S THE IMPACT OF AI ON OUR CAREERS?
How worried you need to be about AI taking your role depends on the industry you work in, or if you’re a student, the industry in which you’re studying to work.
As some industries will see shrinkage, others will see growth, highlighting the complex nature of trying to predict the effects of AI in different industries and sectors.
The financial planners and analysts at Datarails recently highlighted the sectors predicted to be the most negatively impacted by AI are healthcare, finance, and retail. Healthcare will see demand for more roles like data scientists and machine learning engineers but will see job decline among radiologists, pathologists, administrative roles, and laboratory techs.
In finance, the impact of AI will be felt mostly in entry-level roles related to data entry, risk management, and risk assessment. AI’s impact on finance is already very apparent with robo-advisors and programs that streamline financial planning and advising. Retail will be another sector that feels the heat from AI. As shopping becomes less and less of an in-person activity—paired with AI-enabled chat solutions and product recommendations—roles in customer service and support might shrink. On the flip side, industries predicted to be least affected by AI are the arts, human resources, social work, and personal services (like hairstylists, mechanics, and personal trainers, among others).
Statista reports that specific roles across several industries could be automated by AI. Further, the labor market between the years 2023 and 2027 is expected to see a loss of 83 million jobs in tandem with the creation of 69 million new jobs related to AI and automation. I highlight the Statista graphics to clarify that AI is not all bad, and I believe it is a matter of adapting and learning skills that can help us maintain relevance in our careers in this fast-paced, technologically advancing environment.
WHAT IS THE IMPACT OF AI ON THE PLANET?
You’re probably wondering: What are the impacts of AI on the planet? To answer this, first I’ll explain how AI energy use is quantified and break down what AI’s energy consumption is attributed to.
There are two main areas of AI energy consumption. First, there’s AI training, which involves building large models like GPT-4. The Verge reports that training these models requires a lot of computational power. For example, the amount of energy it takes to power 120 U.S. homes for a year is 1,287 megawatt-hours (MWh) of electricity, which is the same amount of energy needed for GPT-3 alone, according to The Verge.
The second area is AI inference, which is the way we all use these models every day like answering queries (for example, how I asked ChatGPT to define AI). While inference uses less energy than training, its impact can still add up. Third Way reports that energy consumption for inference in data centers can vary from a few watts to hundreds of watts per query, depending on the model and hardware. Google’s AI-powered search, for example, has been reported to increase energy usage per query by tenfold compared to traditional search methods, Third Way noted.
Where is this power generated? While there are no dedicated data centers used exclusively for AI to perform its tasks, it relies on specialized electronics housed in existing data centers. There are significant environmental impacts from these data centers—from CO2 emissions to impacts on water—but I’ll explore those in another article.
To shift the focus back to energy consumption, Wells Fargo found that energy demand for AI in the U.S. is expected to hit 652 terawatt-hours (TWh) by 2030. To put this into context, Inspire Clean Energy reports the entire state of New York consumes 143.2 TWh of electricity annually.
It’s important to highlight the effect on energy and jobs by AI is not set in stone. To offer perspective, the University of Michigan has developed an energy optimization framework called Zeus, which can reduce the power needed for generative AI by up to 75 percent.
While we perceive AI to be detrimental to humans and the planet, AI is advancing at a fast pace. We don’t know about all the new careers AI could create and the new power and energy scaling models that could be implemented to lower cost and environmental impact. Even projecting the impact of powering data centers with wind and solar could be difficult.
WHAT’S NEXT?
AI’s influence is vast and ever-evolving, so I’ll be writing a series on the topic and how it impacts you. Part two of this series will explore the origins of the data AI is being fed and how potential biases in the datasets could lead to reinforcement of errors in the models. I will also explore AI centralization and the new AI being created. Lastly, I’ll provide a breakdown of some of the top AI models, their revenue generation, and an overview of the sector.
AI’s potential is immense, but so are its challenges. As we move forward, it’s crucial to approach its development and integration with responsibility, ensuring that its benefits are maximized while minimizing its impact on both people and the planet.
