Experts appear divided, with some arguing there’s no future without AI while many have sounded the alarm over AI’s potential to make a large proportion of human workers obsolete. 

By some estimates, AI innovations will bring a $4.4 trillion boost to the global economy—but it’s also been projected that 300 million jobs will be lost to automation.

But the chair of Britain’s top AI and data science research institute argues the short and long-term impacts of the technology could look quite different.

Douglas Gurr—chairman of London-based AI research center The Alan Turing Institute and former country manager of Amazon U.K.—said Thursday that when new tech upended industries throughout history, it took a long time for society to catch up with all the ground-breaking changes it brought with it.

“Roughly every 100 years, [for the] last three centuries, we’ve seen a transformative technology that has changed the world,” Gurr said, speaking during a panel discussion at London’s CogX Festival on Thursday. “In each of those previous occasions, it took an entire generation [to fully adapt to the tech], and that caused massive social upheaval.”

This was because the existence of most of the jobs people knew how to do was suddenly brought into question, Gurr explained. So, the transition phase of the new technology displaced labor. 

However, once “the dust settled” and the tech’s full potential and impact kicked in, it would uplift society to new levels of wealth, employment and more, according to Gurr. 

“If you just take the world before each of those [transformative] changes, it was a much better world after—the world was wealthier, there’s more money for health, for education, medicine,” he said.

It wasn’t just that there was more wealth after technological transformations, Gurr added, but that wealth was always better distributed. Now, AI came with the promise of making society more prosperous and skilled, he argued. 

In an emailed statement, Mark Girolami, chief scientist at The Alan Turing Institute, told Fortune the emergence of new technologies historically gave rise to “a disruption in the deployment of labor,” with the effect being “a shift in required skills partnered by a rise in socio-economic wealth.”

“Consider the advent of the internet and the world wide web from which the demand for whole new skilled occupations arose such as web developers, network managers, displacing lower-level technical work with these higher skilled occupations,” he said. “It’s critical that we identify and prioritize the development of the right skills in the face of continual advances.”

Concerns the world will ‘repeat mistakes’

Despite touting AI’s potential to deliver positive societal transformation, Gurr admitted on Thursday that he was worried “we will repeat mistakes we made previously” by failing to prepare the workforce for the looming disruption.

“This is exponential, not linear,” the research institute’s boss said of the AI boom. “Linear gradually gets there, exponential [growth] goes from very small to—bang. It goes from quite niche to ubiquitous quite quickly.”

While Gurr said AI could bring about positives like safer transportation via autonomous vehicles and “much, much more income,” he noted that he was worried the world was “behind the curve in terms of the job transition.”

“[If] I’m a lorry driver, this feels like you’re taking my job,” he said. “If we don’t address that, we will repeat [historical] mistakes and it will come with massive social upheaval. We are much better able to communicate at scale now. Clearly what we need to do is say to people, ‘how do we give you the skills you need for this [new] world?'”

Disruptive innovation

Many experts have echoed Gurr’s thoughts on what AI could mean for jobs as we know them—and some organizations have even begun using AI to replace some functions traditionally done by humans. 

Earlier this year, IBM said it would freeze hiring for 7,800 roles that could potentially be done by AI. The tech company’s CEO Arvind Krishna said in an op-ed for Fortune in April that AI offered several benefits to amplify how the company’s employees spend their time.

Meanwhile, billionaire venture capitalist and Silicon Valley veteran Marc Andreessen said earlier this month that a symbiotic relationship between humans and AI could provide “a much better way to live.” 

But not everyone has shared this optimism: Tesla and SpaceX CEO Elon Musk recently called AI a civilizational risk that could hit people “like an asteroid,” while one group of researchers called the tech a “nuclear-level catastrophe.”

Oxford University economist Carl Benedikt Frey, who predicted in 2013 that half of U.S. jobs will be lost to automation, told Fortune in February that OpenAI’s ChatGPT posed a threat of disruption akin to Uber in the taxi industry.    

“The hard part to figure out is how we can use AI to create innovation that then creates new occupations and new industries,” he said.