1 hr ago
Instagram’s 80s Trend Reveals AI Image Generation’s Environmental Cost
People on Instagram are making retro 1980s-style pictures with AI tools.
Each picture requires computers in data centres to do the work.
Research estimates that one AI-generated image can use 0.01–0.29 kWh of electricity.
One image may seem small, but millions of people making several images can use much more energy.
Data centres also use water to cool their powerful computers.
Researchers estimate that 20–50 ChatGPT questions and answers may use about 500 millilitres of water, though the amount varies.
The computers also need minerals and special parts made through mining and manufacturing.
This means a fun online trend can have environmental costs.
The trend may fade, but new AI trends could create similar impacts.
Research estimates that one AI-generated image can consume 0.01–0.29 kWh of electricity.
Millions of users generating several images could create substantial overall electricity demand.
Data centres use significant electricity and water to power and cool AI systems.
Researchers estimate 20–50 ChatGPT questions and answers may correspond to about 500 ml of water use.
Mining minerals and manufacturing GPUs add hardware-related environmental costs to AI image generation.
- Who
- Instagram users, AI systems, data centres, and the researchers from Carnegie Mellon University and Hugging Face are involved.
- What
- A viral Instagram 80s photo trend is prompting concerns about the electricity, water, and hardware resources used to generate AI images.
- Where
- The trend is taking place on Instagram, while the AI processing occurs in data centres; no specific data-centre location is given.
- When
- The trend is currently viral and has only recently started, according to the article.
- Why
- Generating large numbers of AI images requires electricity, cooling water, specialised hardware, minerals, and manufacturing resources.
Key facts
- Electricity per AI image
- Estimated at 0.01–0.29 kWh.
- Water estimate
- About 500 ml for roughly 20–50 ChatGPT questions and answers.
- Main infrastructure
- Data centres with powerful servers and GPUs.
- Cooling requirement
- AI data centres consume water to manage heat from servers and GPUs.
- Hardware materials
- Specialised GPUs and other equipment require minerals, rare-earth elements, and other raw materials.
- Broader supply chain
- The environmental footprint includes mining, semiconductor factories, GPUs, data centres, electricity, and cooling systems.
Sources
Related news
Sam Altman Challenges Claims About AI Water Consumption
ChatGPT 80s Photo Trend Raises Privacy and Environmental Costs
India’s AI Data-Centre Boom Raises Hidden Water Concerns
AI Expansion Puts Jamnagar and Thane Under Water Stress
India’s AI Expansion Faces Growing Power and Infrastructure Demands
AI Data Centres Face Rising Power and Cooling Demands
ChatGPT 80s Photo Trend Sparks Privacy and Energy Questions
India Debates Whether IIT Madras Should Build a Medical School
Vivek Raghavan Named Among TIME’s 100 AI Leaders
