Artificial intelligence could displace between 1m and 3m private sector jobs in the UK, though the ultimate rise in unemployment will be in the low hundreds of thousands as growth in the technology also creates new roles, according to Tony Blair’s thinktank.

Between 60,000 and 275,000 jobs will be displaced every year over a couple of decades at the peak of the disruption, estimates from the Tony Blair Institute (TBI) suggest.

It described the figure as “relatively modest” given the average number of job losses in the UK has run at about 450,000 a year over the past decade. More than 33 million people are employed in the UK.

AI, a technology that can be loosely defined as computer systems performing tasks that typically require human intelligence, has shot up the political agenda after the emergence of the ChatGPT chatbot and other breakthroughs in the field.

TBI added that it did not expect the scale of the displacement to be reflected in long-term job losses. It predicted total losses to be in the low hundreds of thousands at its peak at the end of the next decade as AI creates new demand for workers and pulls them back into the economy.

“Our best guess is that AI’s peak, impact on unemployment is likely to be in the low hundreds of thousands and for the effect to unwind over time,” said the report, titled the Impact of AI on the Labour Market. “A common lesson is that AI is likely to increase the dynamism of the labour market by prompting more workers to leave existing jobs and start new ones.”

Such a process will require an “upgrade” to the UK’s labour market infrastructure, TBI said, which could include an early warning system flagging how a worker’s job could be affected by AI.

The report estimated that deployment of AI could raise GDP – a measure of economic growth – by up to 1% over the next five years, rising to up to 6% by 2035. Unemployment, meanwhile, could rise by 180,000 by 2030. Currently, there are about 1.4 million unemployed people in the UK.

However, TBI, which has described AI as a “substantial policy challenge”, said all of these scenarios are dependent on factors such as what tools emerge over the next decade, investment decisions made by private firms and government policies that accelerate or delay implementation.

The thinktank said AI would “certainly” replace some jobs but could create more by boosting workers’ productivty, which would increase economic growth and create more job openings. It could also follow the pattern of previous technological breakthroughs by creating products and sectors that require workers to perform new tasks and roles.

TBI indicated that administrative and secretarial jobs will be the most exposed to the technology, followed by sales and customer service, and banking and finance. Those jobs will produce the greatest time savings from deploying AI, the report said.

TBI said most of the efficiencies are likely to come from products that perform cognitive tasks – such as chatbots – rather than AI-enabled hardware such as robots that carry out physical work. Thus sectors that involve complex manual work such as construction are likely to be less exposed, it said.

However, jobs involving routine cognitive tasks such as secretarial work are more likely to be affected as well as industries that generate large amounts of data, such as banking and finance, which can train AI models more easily.

TBI said it expected unemployment to rise initially as some companies take advantage of time savings from AI – the report estimates almost a quarter of the time private sector workers spend on their jobs could be saved – by letting staff go.