Project Green Light launched on Tuesday in a UK-first which could signal the future of driving.
Its main aim is to reduce so-called 'stop-and-go' emissions which account for half of emissions at traffic junctions according to Google.
Its Green Light programme uses AI and driving trends from Google Maps to model traffic patterns and make traffic lights more efficient.
Recommendations to change the lights can be implemented in five minutes, and don't require changes to infrastructure. Theoretically it could mean cities can also react to traffic and improve flow, for example if an event or diversion has caused one-way congestion at a junction.
Google has already installed the artificial intelligence technology across 70 junctions in 12 cities including Haifa in Israel, Rio de Janeiro in Brazil and Bangalore in India.
Data from these projects show the potential to reduce vehicle stopping by 30 percent and emissions at junctions by 10 percent, Google said.
Google UK's managing director Debbie Weinstein said: "Already, this early stage AI-powered tool uses existing infrastructure to provide city planners with a cost-effective and efficient way to improve traffic flow, reduce stop-and-go traffic and cut emissions."
Manchester has more than 2,400 traffic lights and a million journeys are made each week.
Transport for Great Manchester's analysis and reporting manager David Atkin said: "With traffic levels now at or beyond pre-pandemic levels, we are working really hard to tackle congestion and are delighted to be amongst the first areas in the world - and the first in the UK - to work with Google on the innovative Green Light initiative," he said.
"The pilot provided valuable insights and teams from both Green Light and TfGM brought expertise and ideas to the table to improve journeys by up to 18% and reduce emissions.
"Our aim is to make the network run as efficiently as possible and we look forward to seeing how we can use what we've learnt from this pilot to improve journey times for all road users."