The Pakistan Tehreek-e-Insaf (PTI) on Sunday launched its election campaign with a virtual rally, which, according to the party, was viewed by over five million people across all social media platforms.

The virtual rally featured an AI-generated statement of founding chairman Imran Khan.

The December 17 online meeting was marred by internet disruptions. However, the telecoms regulator said the interruptions were being investigated, but added that internet accessibility overall appeared to be normal.

The four-minute-long AI-generated speech of Imran Khan thanked his social media team for their efforts. The speech also highlighted that the party “was not allowed to hold rallies and the members were being harassed and kidnapped.”

The statement, based on incarcerated politician’s notes, urged his supporters to come out in large numbers on February 8 and cast votes to change the system.

The party says the voice of their former chairman was AI-generated. The imagery used in the video shared contains both real photos and videos, as well as doctored pictures of Imran Khan showing him behind bars and reading the Holy Quran.

When asked whether the whole video was produced using an AI platform, a spokesperson of the PTI confirmed to Geo.tv that only the voice of Imran Khan was AI-crafted.

Khan, who has been behind bars since August is facing multiple cases including one related to a secret document, which his party claims was proof of a conspiracy to dislodge him.

‘Counter-productive’

Mazhar Abbas, senior political analyst, believes that the party is taking a new direction ahead of polls.

“PTI has hit both, the powers that be and its opponents, with this new narrative of defiance and virtual jalsa is a step in that direction,” he said. He added that the attempts by the authorities to block the virtual rally proved counter-productive.

Michael Kugelman, a scholar of South Asian affairs at the Wilson Center in Washington, noted that the party has large support outside Pakistan and that any moves to disrupt the virtual rally was “wrongheaded.”

Digital Rights Foundation’s (DRF) Executive Director Nighat Dad told Geo.tv the use of AI-generated audio is a “very worrying trend” since many people are unaware of what AI content can do.

She noted that Imran’s “online jalsa” video was not “made sophisticatedly,” but in the lead-up to the polls, the audio may become more sophisticated.

Dad also said the content could be transferred through WhatsApp, where users had no way to identify the pattern of fake news, misinformation, disinformation, deep fakes, and AI-generated content of synthetic media.

“Basically, generative AI is democratising disinformation,” she cautioned, saying that political actors could gain sympathies from their voters with these AI-generated videos and audios to boost their vote bank.