AMERICAN technology giant Meta has introduced, or perhaps one should say, unleashed AI on speakers of the Urdu language. At a Monday event in Islamabad, the tech company said it was expanding its artificial intelligence language capabilities to better serve users in Pakistan who wished to interact with AI in Urdu. The announcement was both exciting and unsettling. “The introduction of Urdu for Meta AI is a milestone in making technology more inclusive and accessible for our people, ensuring that no one is left behind in our digital transformation,” the federal IT minister said on the occasion. While the intent should be appreciated, we must also worry whether there has been enough discourse in Urdu regarding the dangers and shortcomings of AI and how users must protect against them.

There is no denying that AI models that are accessible in ordinary users’ mother tongues can open up whole new worlds of possibilities in a society like Pakistan, where social inequalities often limit who gets to benefit from modern advances and who gets left behind. Technology has always been a great leveller: each innovation has enabled humans to further their individual potential. AI is incredibly more potent than any other technology in this regard because it allows users to connect and engage with it at a very human level. However, as has already been seen, AI has brought with it its own set of dangers. Misuse of AI has already created major challenges, especially in the form of the increased spread of misinformation and disinformation. As it is introduced to new categories of users, there is a risk of contagion, and without adequate safeguards in place, this could create major problems. Therefore, it is hoped that as much effort will be put into educating users who will be interacting with AI in Urdu as has been done for English speakers. They must not be exposed to its many risks unprepared.

Published in Dawn, October 29th, 2025