UAE to launch DeepSeek-inspired AI models

The United Arab Emirates is planning to launch new artificial intelligence models inspired by China's DeepSeek, a senior official told AFP, calling the system's disruptive emergence "fantastic news".

Faisal Al Bannai, the driving force behind the UAE's Falcon large language model, said DeepSeek's challenge to American tech giants showed the field was wide open in the race for AI dominance.

The oil-rich Gulf monarchy is betting big on the transformational technology as part of its push to diversify its economy away from fossil fuels.

Bannai said he was heartened by DeepSeek, a high-performing and apparently low-cost AI model that sent US tech stocks tumbling after its launch.

"It's fantastic news. Because it proves one thing: this game is at its beginning," Bannai said at the World Governments Summit in Dubai on Tuesday.

The UAE launched Falcon in 2023, a large language model that compared favorably with industry leaders including OpenAI's ChatGPT.

"What happened with DeepSeek was another proof that small teams, agile teams, agile countries, can move fast and can make an impact," said Bannai, the UAE's presidential advisor on advanced technologies.

"So we are learning, I think, from what they showed. We are taking learnings, and we will be also launching other models in this regard.

"And I think (DeepSeek gives) a serious sense of encouragement that you can punch way above your weight in this game, because the game is still starting."

Alongside Falcon, the UAE has developed Jais, an Arabic-language AI chatbot, while digitizing and automating government services to a large extent.

It has also been active in physical infrastructure, pledging tens of billions of dollars in recent weeks to build data centers in France and the US.

Bannai, who is also secretary general of the Abu Dhabi-based Advanced Technology Research Council, said the UAE's wealth, ability to attract talent and its top-down decision-making could make it a serious player in AI.

"When the internet phase 1.0 or 2.0 happened, we were not necessarily ready," he said. "Today we are in an amazing situation where we have such a diversified ecosystem as a country over here, talents from all over the place.

"And we are in a fantastic position to move very fast in an age where sometimes people are taking time to decide... this is a time where countries or companies that can make good decisions... can move fast."

He was speaking on the same day that US Vice President JD Vance warned world leaders meeting in Paris against "excessive regulation" of AI.

The United States and Britain also refused to sign a statement calling for regulation to ensure the technology was "open" and "ethical".

"Trying to assume any country can regulate AI, I think is delusional, because if you regulate, someone else is not regulating in the same way," Bannai said.

"When it comes to AI, I think regulation needs to start focusing on a few areas that matter but doesn't stifle the growth," he added.

"Because, frankly, whoever puts the regulation, the other guy might be more practical, and life will move on."