KARACHI: Without developing domestic AI capabilities, Pakistan may become increasingly dependent on foreign AI solutions, Byungjo Kong, Digital Technology Specialist (Data Analytics and Big Data) at the Asian Development Bank (ADB), said in response to questions via email.
This dependency not only leads to higher costs but also places critical infrastructure and systems under the control of foreign entities, creating potential economic and security vulnerabilities, he explained.
“AI should be built to address a country's specific needs and challenges,” said Fatima Yasmin, vice president for sectors and themes at the Asian Development Bank. The first wave of AI could widen disparities between countries, as is happening in Pakistan, which has lagged behind other countries in adopting AI.
Lessons from India
In India, AI is being used for a variety of purposes. For example, in Tamil Nadu, a military-grade AI surveillance system is being used to prevent elephant deaths on railway tracks. AI cameras have been installed on 12 towers along two major railway tracks, which alert forest and railway officials if an elephant is within 100 feet of the tracks. So far, nearly 400 cases of elephants approaching the tracks have been detected. Given this success, Tamil Nadu plans to expand the AI-based system to other risk areas in the state.
Another example is Kissan GPT in India, a chatbot introduced to help farmers irrespective of their literacy level. The chatbot allows farmers to ask questions on their mobile phones and it understands languages like Hindi, Gujarati, Marathi etc. and responds with detailed answers.
“India will be a major player both as a producer and consumer of AI,” said Ozer Khan, director of digital innovation and architecture at the Asian Development Bank's Information Technology Department. Such uses of AI could also be adopted by Pakistan.
OLAM International, a leading agribusiness company, is using AI to streamline rice processing and distribution in the Philippines. By simply uploading a photo of the crop, the AI assesses the quality and connects farmers with buyers, he said. Such applications could help address Pakistan's arti problem.
Singapore has purchased S$500 million worth of graphics processing units, which are essential for developing AI apps, but they are currently unused, he added. [Singapore] “They are hoarding AI. In Pakistan, where there are challenges, is the government making such AI investments?”
Widening gap
Kong detailed the impact of the widening AI divide, which could be especially significant for countries that have not yet prioritized AI investments. AI has the potential to significantly boost economic growth and productivity. AI can drive innovation and competitiveness by automating processes and increasing efficiency in manufacturing, healthcare and services.
Pakistan faces a talent dilemma. Kong says countries with strong AI strategies have created education and professional development programs to train AI experts. Pakistan's lack of such efforts means it's not only missing an opportunity to improve its own tech capabilities, it's also facing a talent drain, further reducing the country's chances of catching up.
It is important that governments seriously consider these implications and start incorporating AI strategies into national development plans, which will include not only investing in the technology but also building the infrastructure and education systems needed to support these advances, he added.
Avoid “copy and paste”
However, he warns that adopting “copy-and-paste” AI techniques that are not adapted to Pakistan's unique needs carries various risks and could be problematic in multiple ways.
Imported AI solutions may not address local languages, infrastructure constraints, or cultural nuances, leading to inefficiencies and bias. For example, AI technologies developed for other regions may not be able to effectively address Pakistan's diverse linguistic and cultural landscape, preventing certain demographic groups from benefiting from AI advancements.
Another potential issue is cultural differences between the West and the East: implementing AI systems that are not designed with local laws and social norms in mind poses significant risks related to data privacy and security.
These challenges extend to the agricultural sector: If Pakistan imports an AI-based agricultural monitoring system designed for a different climate or crop type, it could produce inaccurate predictions about the country's agricultural environment, potentially leading to misguided government intervention and economic losses.
To mitigate these risks, Conn suggested that Pakistan should develop its own AI solutions or tweak imported solutions to meet its specific needs. Investing in local research, establishing partnerships, and creating a national AI strategy could help ensure that technology works with local relevance while strengthening domestic capabilities.
In an era dominated by technological advancements, AI stands out as a force that will transform industries and societies. With India rapidly emerging as a major player, Pakistan finds itself at a critical crossroads. “Without a national framework and strategy on AI, Pakistan stands to miss out on opportunities as well as the means to keep its people safe,” said Khan.
Published in Dawn on May 26, 2024
