Expert Weighs In On Environmental Impact Of Deepfake Videos That Is "Often Overlooked"

Deepfake technology creates false information in the form of images, videos or audio with the help of artificial intelligence (AI). These media look real, but are actually fake. The rise of deepfakes sparked concerns over risks of unethical use. Meanwhile, a recent report highlighted another issue as well: the environmental impact, BBC News reported. 

Dr Kevin Grecksch, who is a lecturer at the University of Oxford, said that the videos are generated using tools like OpenAI's Sora, which require massive computational power and energy to operate.

The process of creating these videos typically occurs in large data centres, which need a huge amount of electricity and water for cooling purposes.

Also Read | Police Detain US Teenager After AI System Mistakes Bag Of Chips For Gun

He said that there was "quite a huge hidden impact on the environment" because "somewhere these videos have to be produced - and that's typically not on your phone".

"It happens in a data centre which could be somewhere across the world, or maybe just around the corner, you never know," Dr Grecksch said as quoted by the report. "First of all that uses up a lot of electricity and secondly it uses a lot of water."

Dr Grecksch emphasises the need for integrated thinking regarding the location and operation of data centres, especially in regions marked for AI growth.

"There's a lot of water involved, and I think we just need to think about what we're using it for, how we're using it, and how often we're using it," he said.

It is important to consider the environmental impact of deepfakes, which is "often overlooked".

Also Read | Telegram CEO Pavel Durov Offers To Buy Stolen Louvre Jewels, But Not For Paris

The popularity of deepfake apps like Sora has increased recently on social media platforms. Notably, Sora has dominated the Apple App Store charts in the US with over a million downloads in less than five days.

"The government sees south Oxfordshire as one of the first AI growth areas, but that's a massive problem because apparently no one has thought where the water is going to come from that's going to cool down those servers."

"There's a lot of questions we need to think about."