The world’s global greenhouse-gas emissions could be cut significantly after Google DeepMind began to implement AI in the search giant’s data centres.

Overall energy consumption at the huge buildings which house the servers containing internet users’ historical data is expected to be reduced by 15 per cent.

"Being able to put a dent in that benefits the world in general," said DeepMind co-founder Mustafa Suleyman.

Data centres are power hungry because energy is required to keep the equipment cool – and some research has suggested they are responsible for two per cent of total greenhouse-gas emissions.

The machine-learning algorithm cut the cost of cooling by 40 per cent, he added.

"This will have a significant environmental impact," he said, before confirming that the technology will be rolled out to all of Google’s data centres by the end of the year.

Google has 16 data farms - nine in the US, five in Europe and two in Asia.

And the tech will not be confined to Google’s centres, with a white paper expected in six weeks.

"We are already talking to non-Google partners about using the same algorithm," Suleyman said.

Meanwhile DeepMind is seeking to win the public over to its controversial NHS partnership with a series of patient meet-ups.

British start-up DeepMind Technologies was bought by Google in 2014 and renamed Google DeepMind.

Its AI –monikered AlphaGo – famously beat Go grandmaster Lee Sedol 4-1 at the ancient Chinese board game.

It was also ‘taught’ to play a retro Atari video game.