Chief AI researcher in UAE says focus must be on solutions, not just cost, after DeepSeek

Developing platforms addressing real-world problems is more important, says the Technology Innovation Institute’s Hakim Hacid

Dr Hakim Hacid, chief researcher at Abu Dhabi's Technology Innovation Institute. Photo: TII

The emergence of more affordable artificial intelligence platform DeepSeek has reshaped the world view of the cost of developing AI, and overshadows the greater need to address real-world challenges, the chief researcher at Abu Dhabi's Technology Innovation Institute said on Wednesday.

While pricing is a factor governing which tier of service users get, more focus is needed on developing AI applications that solve problems, rather than merely identifying them, Hakim Hacid said during a session at the AI Everything summit in Dubai.

While acknowledging that DeepSeek – which is, depending on the user tier, free or significantly lower priced than established models such as ChatGPT – gives more people AI access, he cautioned that the situation may be “a little bit exaggerated”.

Dr Hakim Hacid, chief researcher at Abu Dhabi's Technology Innovation Institute, left, speaking to Salim Essaid, business editor of The National, at the AI Everything summit in Dubai on Wednesday. Photo: Supplied

“When it comes to the cost, we need to be careful on what we are actually looking at so we understand that this cost is not the cost of the model as a whole – it's [just] part of the model,” he said.

DeepSeek's recent arrival has upended the market. The significantly more affordable model has heightened the competition in the generative AI scene. However, its appeal will fade quickly if its more practical uses, especially to solve problems, do not receive sufficient attention.

“We believe that, in a couple of years, these things will become a commodity,” Mr Hacid said. “The AI that we have today is good on the leaderboards, but it's not necessarily good when it comes to the applications [to problem-solve].”

DeepSeek and ChatGPT: how do they compare?

01:09

In terms of the types of solutions, Mr Hacid said attention should be directed to advanced-use cases such as in key industries that serve people and businesses.

“We are used to seeing these chatbots; everybody is excited about them. But I think we have passed that period. Now we are focusing on more complex problems,” he said.

Among the sectors poised to harness AI innovations are food security, biotechnology, health, engineering and several science fields, which could all contribute to developing a full-fledged smart city, as the UAE envisions.

The Technological Innovation Institute is working on “a lot” more AI projects, Mr Hacid said. One of them is the creation of a start-up focused on agriculture using satellite imaging, desertification methods and rain patterns, among others, to address food security.

“We're able to predict a certain amount of phenomena that were not able to predict before,” he said.

The initiatives in AI are helping propel the UAE higher on the global stage, Mr Hacid said.

“The UAE was among the first countries that put a bet on AI. We are not yet at the level of [other economies], probably because of the [difference between economic] sizes, but there is clearly a very good position and very important role that UAE is playing,” he said.

“The region and the UAE don't want to be a consumer of AI any more, but they want to be an active player where they can contribute and where they can have their voice as … an adviser and active participant in the domain.”

The region and the UAE don't want to be a consumer of AI anymore, but they want to be an active player where they can contribute and where they can have their voice as an adviser and active participant in the domain

Hakim Hacid,
chief researcher at Abu Dhabi's Technological Innovation Institute

Mr Hacid also said the institute is ramping up its readiness on quantum computing, which he believes will be a critical part of technology's future and would complement AI. However, there are several issues to address before it becomes mainstream.

“There is a lot of work that is currently done … around generative AI and quantum computing in terms of risks and … security aspects,” he said.

“We see signals coming from this current AI [technology], but we still have some work to be done to be able to make sure that we are really contributing to solving the problems instead of just surfacing those problems.”