Think Twice Before Reaching for AI: The Hidden Environmental Cost
Box Bear Ltd
Where media evolves: Innovative technology and creativity combined to excite your audience
Stop! Yes, you—the one about to click 'generate' to create a Van Gogh-stylized portrait of your cat. Did you know that a single AI-generated image can consume as much energy as half of a smartphone’s battery? While it may seem like a harmless digital task, the environmental toll is far from insignificant.
While digital services like Bitcoin mining and NFTs (Non-fungible tokens) have become infamous for their environmental impact, AI is quickly joining the ranks. This issue was recently explored in an episode of the Artificial Human, How Green is my AI, where hosts Aleks Krotoski and Kevin Fong discussed the significant carbon footprint of data centers. These facilities, which provide the immense energy needed to power AI models, now have a carbon footprint that rivals that of the aviation industry.
And there’s no sign of slowing down. Energy consumption by data centre's is projected to double by 2026, with AI being a significant driver. As we lean more heavily into automation, machine learning, and AI-generated content, many users remain unaware their casual use of these technologies could be undermining global efforts toward net-zero carbon emissions.
At the risk of oversimplifying, there’s an important conversation to be had about digital democracy. While AI offers incredible opportunities for innovation and efficiency, it’s crucial that we also consider its environmental cost. By making more informed decisions about our digital habits, we can contribute to a more sustainable future.
领英推荐
At Box Bear, we take this into consideration when we make decisions surrounding AI. We understand the value that AI offers, but also insure that we have have a strong rationale for when we use it. We don’t take that decision lightly: We use artificial intelligence, intelligently.
So, before you click 'generate,' think twice. The hidden cost of AI might be more than you’re willing to pay.
Head of the pharma team at Box Bear Ltd - innovative digital solutions for effective communication
5 个月Aside from the environmental impact, there’s also the issue that using Ai for search is also prone to giving wrong answers- and demonstrably so. Thats because the algorithm just makes stuff up based upon probability. See the recent examples of asking a LLM how many of the letters “r” does the word strawberry contain. Or a case in medical imaging where the Ai was meant to be diagnosing lung cancers. But instead it had just learned the name of a hospital in the corner of the scan file as giving the “right” answer.