The era of artificial intelligence (AI) technology in political campaigns has officially begun in Pakistan as the country prepares for upcoming elections.
In the first-of-its-kind event in Pakistan, the country’s former prime minister Imran Khan, who is behind bars, delivered a speech using AI to create a voice clone to address a virtual rally.
The voice clone allowed the jailed politician to campaign for his party months ahead of an expected general election.
Khan has been locked up since August 2023 after he was found guilty of corruption. He is additionally accused of leaking classified documents. The cricketer-turned-politician has repeatedly denied allegations intended to stop him from contesting nationwide polls expected in February 2024.
First use of AI in Pakistani politics
The 71-year-old leader’s political party released a four-minute video on Sunday (December 17) evening that used AI-voice cloning technology to replicate his voice. In the video, which aired during a “virtual rally” in Pakistan, the dubbed audio is accompanied by a caption that states, “AI voice of Imran Khan based on his notes.”
AFP reported that Khan sent PTI a shorthand script, through lawyers, which was later edited by a legal team to better resemble the politician’s rhetorical style. The resulting text was then dubbed into audio using software from the AI company ElevenLabs, which makes a text-to-speech tool and an AI voice generator.
“This was a no-brainer for us, when Imran Khan is no longer there to actually meet at a political rally,” which is why the team came up with this idea to reach out to the masses and get ready for elections, PTI social media lead Jibran Ilyas told AFP. “We wanted to get in election mode,” Ilyas said. “No PTI political rally is complete without Imran Khan.”
PTI social media focal person Azhar Mashwani told Dawn that the content of the video is based on notes provided by Khan from prison, adding that the AI-generated voice feels like a “65-70%” match of the former prime minister.
The audio was broadcast at the end of a five-hour live stream of speeches by PTI supporters on different social media platforms.PTI said the event was viewed by six million people across YouTube, Facebook and X.
Voicebot.ai reported that Khan’s social media team used archival footage from Khan’s older rallies and ElevenLab’s synthetic audio engine to clone the politician’s voice.
The PTI was the first political party in Pakistan to widely harness the potential of social media, using apps to target younger audiences who carried them to power five years ago.
The use of AI voice in Pakistani politics has highighted the potential use of AI around the world as polls are scheduled in the United States, European Union, Russia, Taiwan and beyond in 2024.