The UAE's ambition in artificial intelligence clashes with the reality of American and Chinese dominance

Within months of the launch of the Chatgpt app at the end of 2022, research labs in the UAE announced the development of reliable models for artificial intelligence. The list is issued when the Falcon model, an open source system that is widely popular, was developed with government support, along with the “Jais” model, named after the highest mountain peak in the country. In a statement released in 2023 after the launch of the “Jess” model, Ping Xiao, CEO of the Emirati Technology Group “G42” (G42), said: “The future of artificial intelligence is no longer a dream, but rather a tangible reality.” But the current reality indicates that the Emirati dream to build artificial intelligence models is still far. The Falcon model shows a clear delay compared to the advanced models developed by US businesses, either in terms of the number of users or general classifications. Meanwhile, G42 recently withdrew its resources from the “Jess” project, and rather focused on developing customized features based on models developed by other businesses, such as “openiai”. Xiao told Bloomberg in April: “In the early days we were not aware of what the extent to which the basic models could develop or what needed to reach the next phase … It is not logical for a country that has our volume to follow this path,” Xiao told Bloomberg in April. The global race increases more than two years after the introduction of the Gulf of obstetrician artificial intelligence between Washington and Beijing. The world race has become to develop more sophisticated models that are increasingly limited to only two main forces. While a handful of US businesses lead this race by pumping billions of rands on electronic chips and data centers and attracting talents to build the largest and strongest models, China accelerates to catch the knees, and the market is plunged into an open and low cost models to improve its technological influence. As far as most other countries are concerned, even rich, such as the Emirates, they find themselves ‘lost’ in the middle of this race. In this context, there were many artificial intelligence projects that were previously promising in the Middle East, and Europe was a sharp decline or the development of continued development quit completely. For example, the German company “Alph Alpha”, which was previously sung as a European participant to “oben ai” obence ai “, made a decision as the company” G42 “made last year. Regarding the British company “Stability Ai”, one of the pioneers of the early models, it is largely relaxed due to administrative problems. Even the French company “Mistral”, which received great support from investment capital supported by the country’s government, has not yet had strong indications of real commercial success or wide interest from developers. Restore artificial intelligence strategy in the region as in many global markets, restoring businesses in the midste the feasibility of investing in building advanced artificial intelligence models of scratch. Kirill Evitimov, executive director of the Core 42 (Core 42), indicated that the “Jess” model could be an ‘advanced and competitive’ model if the company still invests in it. He added: “But is it the right commercial strategy to utilize a market share? Most likely no.” In 2023, the UAE launched a new company called “A71” (AI71), which is described as a commercial arm of the “Falcon” model, developed by a government research party. However, AI71 eventually used a similar approach as ‘G42’ and ‘thousand thousand’, and focused its efforts on the development of artificial intelligence instruments designed specifically for specific commercial uses, depending on different models, including ‘Falcon’, according to a company’s spokesman. “Falcon” is facing global competition, although “Falcon” is still the most competitive model at Emirate level, but it suffers from the problems of keeping up with the accelerated developments introduced by Open Source models by companies such as “Meta platforms” and “DeepSeek” Chinese. In 2023, the Institute of Technology Innovation (TII), the developer of “Falcon”, announced that the model first place under Open Source models on the Hugging Face platform, one of the most important classification indicators in the artificial intelligence sector. However, the data, which begins last week, showed that “Falcon” is no longer on the list of the best 500 models on the platform classification panel. A representative of the Institute said the “Falcon” exceeded the 55 million drop off. However, this number is still limited compared to Llama models of “Meta”, which won more than a billion downloads. He also pointed out expected plans to integrate the model with companies working in cyber security, robots and cloud computing, which will soon be announced. He emphasized in his statement that “the value of the artificial intelligence model is not only measured by the results of performance in the initial tests, but also in the ability to support and expand innovation over time.” Attempts to break American and Chinese hegemony have G42 on its way to other ways to take advantage of the current mutation of artificial intelligence away from the development of models from scratch. The company has expanded its activities in the data centers in the Gulf region, and the MGX Investment Fund, which it participated in the establishment, provided support for artificial intelligence developers in the United States, including the “Oben AI and Xai” businesses. Countries seek to develop the so -called sovereign models to control more in the way to train artificial intelligence systems, either to direct the way they work, or to reduce the dependence on foreign technology, according to Michael Brunstein, Professor of Artificial Intelligence at the University of Oxford, during the conference “Machines Can See” in Dubai. But he warned that the spread of Open Source models made it difficult to maintain the competitiveness of these initiatives. At the same time, some of these models were also presented as a way to represent languages ​​and populations that are not adequately represented in Silicon Valley models. For example, the ‘oben ai’ and its peers that are largely trained on the internet data, which clearly tend to the English content. But Tarjama CEO Nour Al -Hassan, who specializes in Dubai -based translation services, said that the development of an effective tool in Arabic should not necessarily build a major linguistic model of scratch. She explained that the Arabic.AI Company, a new business offered by ‘Translation’, recently launched an artificial intelligence system developed by an accurate control of a group of large models, an approach you see is more effective for businesses that attach importance to costs in their services. She added: “I don’t need large models of this size to perform a specific task for a bank, for example.” “Jess” is beyond the priorities of “G42”, the “Jess” model is designed by the company “G42” to be dedicated to the operation of Chat Robots in Arabic. In September, the company introduced a linguistic model in Indian under the name “Nanda” as part of its expansion strategy within the Indian market. Incevtion director Andrew Jackson said the company and its research partners will continue to work on updating these two models. However, he also added that “this is not the commercial focus for us. We have to make money.”