Middle East unlikely to slow down on AI, ChatGPT rollout

While an open letter is calling for a global pause in the development of advanced AI (artificial intelligence), with thousands of signatories including Elon Musk and Apple co-founder Steve Wozniak, the Middle East wants to fast forward its development with its own criteria and safeguards. 

On Wednesday, Astra Tech, backed by Abu Dhabi AI company G42, said it created the first Arabic language AI chatbot.

What happened: The Future of Life Institute (FLI), an American nonprofit working to reduce global technology-related catastrophes since 2014, created an open letter in late March calling on all AI labs worldwide to pause "training of AI systems more powerful than GPT-4" for six months.  

The systems it's referencing are mainly language learning models (LLM) that can think, learn and behave like humans that are being considered a potential risk to humanity. ChatGPT-4 is the latest such LLM released by research laboratory OpenAI and was launched last month.

By simply typing any question in a chat box, the AI tool will respond — and is even capable of writing academic reports or poetic sonnets — and if allowed, answer questions such as how to build a bomb.

Given their ability to answer questions like a human, LLMs have the potential to replace low- and middle-skill jobs. It can also inject misinformation given that, like humans, its answers are sometimes misleading, biased or simply untrue. 

In order to generate these answers, it tracks personal data extracted from the web, or parameters, to train its systems. This has created fears about the lack of regulation over data privacy for this rapidly growing technology first introduced late last year.   

Middle East take: The number of signatories of the open letter grew about sixfold since last Monday’s 3,000, highlighting the growing concern about unregulated AI. Among the signatories are names from across the Middle East and North Africa (MENA) such as data scientist Fatima Zahra Mouhsini in Morocco, retired Saudi government general manager Suleiman Mohammed A Aleisa and UAE-based artist May Ali. 

Yet the region is determined to push AI use forward, particularly the UAE and Saudi Arabia, to grow their economies and create their own AI solutions to region-specific issues.    

The UAE has set an AI strategy to become one of the world leaders in AI by 2031, with a digital economy that contributes 20% to its gross domestic product. It has a minister of state for AI, and the Mohamed Bin Zayed University of AI is dedicated to building the nation’s knowledge economy in this field. 

Saudi Arabia launched its National Strategy for Data and AI in 2020, which seeks to attract $20 billion in foreign and local investments by 2030. It’s also building megacities like NEOM and The Line, for which AI is called the “beating heart.” 

MENA applications: Moe Abeidat, group vice president of technology at Aramex, told Al-Monitor that the region is not pioneering new science in AI or ways to train algorithms yet compared to the greater global industry, but given its high AI adoption rate, it needs to be a part of the global conversation. 

Especially when it comes to language processing in Arabic, he said, “No one is going to solve that outside of the region” and pioneers are needed. One such contender is Astra Tech, the owner of voice-calling app Botim, which said it established the first "Arabic ChatGPT" last week in its pilot test phase. 

Abeidat said AI is being used by Aramex to address a MENA-specific challenge of non-definable, or “fuzzy,” delivery addresses. 

“We’ve invested very heavily to use AI to decipher what people put in as addresses to determine precise latitude and longitude,” Abeidat explained. He told Al-Monitor that Aramex is welcoming ChatGPT to improve on this process, as it plans to gradually incorporate it into its internal system as an additional aid. 

“We're leveraging advanced AI to utilize it in a cultural context that benefits us,” said Abeidat, who is also a non-executive adviser with overwrite.ai, a Dubai-based proprietary foundation language model that generates automated and original descriptions for real estate advertisements. 

Future risks: Generative AI is the term used for algorithms like ChatGPT that can be used to create new content that didn’t exist before, including audio, code, text, videos and images. 

For example, it can generate creative videos of a unicorn galloping along the rings of the planet Saturn. If prompted, it could be asked to generate an image of what the Prophet Mohammed would look like, which is forbidden in Islam and would likely cause societal uproar as it did with the French satirical magazine Charlie Hebdo in 2011 and Danish cartoonist Kurt Westergaard in 2005. 

Dimitry Mihaylov, chief scientific officer at metaverse gaming platform Farcana, said the Gulf Cooperation Council (GCC) is equipped to handle unprecedented issues given its ethical guidelines, well-managed firewalls and top-down streamlined governance. 

“They can create their own regulations and can implement them instantly. Like during the Qatar World Cup when they decided to ban alcohol in stadiums, it was done in one day,” he said.  

AI replacing humans: Mihaylov also believes that FLI’s open letter has created heightened panic around the idea of losing jobs to LLMs and generative AI, which he said the Gulf is not afraid of. 

“Actually, it will attract more high-level talent who can handle generative AI and reduce the number of low-skilled workers in the region,” he said, which is important to GCC countries as expatriates make up over half of their labor force and as much as 90% in some, according to the International Monetary Fund. 

Among these GCC populations, there is a shortage of specialized skills in Kuwait (75%), Qatar (60%), Saudi Arabia (46%) and the UAE (46%), according to the PwC's 2022 Middle East Workforce Hopes and Fears Survey. 

Conclusion: Although Italy became the first Western country to ban ChatGPT last week, it’s highly unlikely that a Middle East country, especially one within the GCC, would follow given the region's economic interests and top-down regulatory frameworks. It’s quite the opposite. Dubai Electric and Water Authority was eager to announce in February that it will be the first UAE utility company to use the conversational AI agent to improve its offerings. 

According to Bloomberg, Abu Dhabi AI company G42 hired a former executive from Chinese e-commerce giant JD.com Inc. to manage its $10 billion tech fund and expand the company’s footprint across Asia. 

If there is a pause in AI, said Mihaylov, “It can be an opportune time for this company (G42) to use the pause and grab the best assets, similar to [the] Saudi’s sandbox.” 

The Saudi Data and AI Authority (SDAIA) launched its Data and Privacy Regulatory Sandbox in early February to capitalize on the power of data all while creating protective measures for consumer rights.

In a 2022 IPSOS global survey, Saudis ranked second (76%) in seeing that AI has more benefits than drawbacks. Chinese respondents ranked first (78%) while Americans (35%) were amongst the lowest, anticipating future AI failures.