Lisbon, Dec. 2, 2025 (Lusa) - In an interview with Lusa, academic and AI policy expert Ramayya Krishnan said that in the short term, the greatest opportunities lie in building artificial intelligence infrastructure, such as data centres and gigafactories.
Ramayya Krishnan was one of the speakers at the Responsible AI Forum 2025, which took place on 25 November in Lisbon, where he addressed the topic of artificial intelligence (AI) and the future of work.
Asked how he sees the future of work with AI, given that technology is evolving at a rapid pace, the professor and director of the AI Measurement Science & Engineering Centre at Carnegie Mellon University (CMU) said that, "at one level," artificial intelligence "is developing rapidly."
"To be honest, I think that in the short term I see the greatest opportunity really in building infrastructure for AI," i.e., data centres and gigafactories, "things of that nature," said the official, who was a member of the US National AI Advisory Committee.
These are jobs "in construction, skilled jobs, which are about how these data centres are actually built," he continues.
Generally, "they take between 18 and 24 months, and it's a lot of jobs," because electricians and construction workers are needed.
In other words, "many skills and qualified people are needed. In fact, in the US, this is a major concern: we don't have enough people to carry out the huge infrastructure construction underway. So I think that's in the short term," he adds.
"In the medium term," between three and five years — because "once these data centres are built, the number of people needed to manage them is very small; it takes a lot of people to build them, but very few to manage them" — "it is necessary to have a strategy for what to do with the data centres and with the AI" that is in these infrastructures, he said.
"That's why we need to build downstream translational centres (...), which actually enable small and medium-sized enterprises, other businesses and the public sector to adopt AI, thus enabling new types of work to emerge," with people having to be trained for this type of work, adds Ramayya Krishnan.
Now, "how this will unfold is much less clear, the first part is much clearer," he considers.
"I can't say that in five years it will be this or that because I think things will probably change. How exactly they will change, we will find out," he reflects.
The first signs are, for example, in software development, "we are seeing the implementation of co-pilots; in customer service, we are seeing the implementation of AI" and "this has not yet led to a very significant reduction in employment," he says.
Therefore, "it's not as if AI is going to be implemented and many people are going to lose their jobs," that "is not yet clear."
In short, "there will be a big increase in infrastructure construction." Then, how exactly AI will be integrated and implemented: will it increase human work or replace human work, will it make a difference in terms of exactly what series of skills people will need, are some of the questions.
"Once AI is used, it does not mean that the amount of work has to remain fixed. You can increase the pie so that it may require more people to do the work," he says.
ALU/ADB // ADB.
Lusa
