AI's Energy Usage May Be The Death Of Us
It's certainly not helping us combat climate change. But apparently it is helping us with grief.
Generative AI is now part of our daily online experience, but energy consumption is spiraling out of control. This hyper-consumption era, fueled by AI's resource-hungry processes, stresses local power grids and causing massive water evaporation.
AI tools are now being integrated into everything, and I mean everything. AI is embedded into our online interactions, from Google searches to Facebook prompts. But the computing power needed for generative AI is exponentially greater than that needed for traditional services. Researchers estimate these applications are 100 to 1,000 times more computationally intensive.
Google's energy consumption doubled from 2019 to 2023, primarily due to the AI race. Despite Google’s claim of not spiking energy use specifically for AI, their total emissions surged nearly 50% compared to 2019.
Last month, Bill Gates urged environmentalists and governments to "not go overboard" on concerns about the huge amounts of power required to run new generative AI systems. Yet, Microsoft faces similar challenges, with its total carbon emissions rising almost 30% since 2020 due to tens of billions of dollars invested into data center construction.
AI’s water consumption is equally alarming. Data centers evaporate millions of gallons of water, which doesn’t return to the earth for up to a year. This evaporation contributes significantly to local environmental stress, unlike residential water use that typically cycles back through sewage systems.
Now,even bitcoin miners are pivoting to AI, converting their energy-intensive mining centers into AI computing hubs. A shift that will only further escalate the energy crisis. Bernstein predicts bitcoin miners’ power capacity dedicated to AI will increase to 12 gigawatts by 2027, doubling their current supply.
Tech companies often position AI as part of the climate solution, but their energy consumption tells a different story. Efficient hardware and smaller AI models could mitigate some impact, yet AI's rapid growth and integration into all aspects of tech threaten to outpace these solutions.
But can we outpace the current climate crisis?
In other news:
If AI wan’t already doing enough, or being talked about enough for your liking, now Netflix teases a trailer for a “griefbot.” A film named “Wonderland” is set to premier soon on Netflix about a company who launches an AI app that lets you talk to your lost loved ones.
It’s making those Black Mirror episodes feel all the more real. While this is just a fiction film, it brings up the conversation on how close we actually are to this becoming a reality. Actually, it’s already happening.
In China, AI-generated avatars of deceased loved ones are booming, offering a new way to “process grief.” Companies like Silicon Intelligence create digital replicas that can converse with users, blending tradition with cutting-edge technology. Despite ethical and technical challenges, the demand is high, driven by cultural practices of honoring the dead. This business model is not only thriving but becoming more affordable, making "digital immortality" accessible to the general public. While the topic of “digital immortality” is a whole other thing to discuss, the impact on mental health and ethical considerations remains debated.
While AI is definitely not helping us with climate change initiatives, it’s apparently helping us deal with our grief. What are your thoughts on this? Are you for or against this? Is AI going too far?