THIS is with reference to the report “Disruptions mar PTI’s ‘virtual power show’” (Dec 18). The way the event was given coverage in international media would have convinced those who had not watched or listened to the cloned act that some political milestone has been achieved. It was not the case, though. It was all based merely on some text-to-speech software that was most likely generated for Hindi audience instead of Urdu speakers. If in doubt, just watch the video and note the pronunciation of letters that are spoken differently in the two languages.

What made me worry a lot more than the political aspect of it was the great disservice such gimmickry did to the country’s technology-related prospects. The gut reaction of governments in Pakistan, regardless of who happens to be heading it at any point in time, is to ban the technology, or at least disrupt the momentum.

As such, believing in the hype that is generated by the party and its followers will help the authorities in completely misunderstanding the limitations of the available technology.

The technology has not advanced to a level where photoreal motion capture and high-fidelity audio can generate video clips of individuals that may mimic real content. Mercifully, the world of deep fakes is not there yet as far as Pakistan is concerned.

Only a handful of companies have the resources and the interest to develop photoreal and hi-fi simulation. Facebook parent company Meta is one of them. An interview of Mark Zuckerberg a couple of months ago in the Metaverse is widely available on social media for a reality check. It is very close to the real deal, but we are not there yet. In essence, long conversations require maintaining the parameters at a consistent level, and, for simulations, that means you can tell pretty quickly that it is an animation and not the real thing.

Of course, there is scepticism. Noam Chomsky made a mistake when he called ChatGPT nothing but “glorified autofill” and “sophisticated, hi-tech plagiarism”. Renowned scientist Michio Kaku also called it an act of “basically plagiarism using digital computers”. Unfortunately, none of these explanations takes into account the resemblance of the artificial neural networks to the human mind or the human-like training process they undergo. The text that is generated by the chatbots is human-like, not because it is plagiarised, but because a human-like mind generates it. It is a far more powerful mind, but it is constructed on the same principle.

It must not be forgotten that the real challenge lies in the tech displacement of the workforce. While the scale might be different, before our eyes, smartphones have displaced many businesses, like video stores, bookstores, music record shops, and photography shops. A global discussion is going on about the future of jobs. Sadly, Pakistan is not even a part of such a discussion. As they say, if you are not at the table, probably you are on the menu. This is what makes me worried. More misunderstanding will lead to more restrictions, which, in turn, will lead to even lesser chance for Pakistan to be ‘at the table’.

The stakes for the country are too high to allow any gimmickry — political or otherwise — to continuously cause harm to the country’s technological land-scape, which is already not in a great shape compared to the global scenario.

Farrukh Pitafi
Islamabad

Published in Dawn, January 14th, 2024