ECP decides to formally use AI in elections, bans deepfakes and makes labels mandatory

ISLAMABAD: The Election Commission of Pakistan has decided to formally use artificial intelligence (AI) in the election process. It has proposed amendments on AI to the codes of conduct for political parties, the media and observers. It has also sought suggestions from the public and political parties to prevent the misuse of AI in the election process.

For political parties and candidates

The proposed code of conduct on AI has been issued to political parties. Under the proposed amendment, political parties, candidates and supporters will not be able to publish election material created or altered by AI without a clear label. Deepfakes, synthetic voices and videos must be clearly marked as AI-generated. Creating and publishing AI content that deceives voters will be prohibited.

The proposed amendment says that forging a person’s likeness, voice or statement will be strictly prohibited. Fake deepfake content against judges, election officials and the armed forces will be banned. Creating fake trends and likes through AI bots and automated accounts will also be banned.

Robocalls and voice-clone messages must state at the start that they are AI-generated. All AI content must be recorded and provided to the Election Commission.

For the media and observers

According to a private television channel, the Election Commission has also issued proposed amendments on the use of AI by the media. Under them, all print, electronic and digital media and social media influencers will clearly label AI content. The media must disclose the use of AI when broadcasting an AI image, video or audio on a candidate or an election matter. Creating fake, derogatory and defamatory deepfake content will be strictly prohibited.

Media houses will be able to use AI only for content generation, translation and analysis. The Pakistan Electronic Media Regulatory Authority (Pemra), the Sindh Media Production Regulatory Authority (SMPRA) and the National Cyber Crime Investigation Agency (NCCIA) will monitor AI content. Action will be taken against unlabelled AI content, and a report will be sent to the Election Commission. On the Election Commission’s direction, the media will be bound to remove misleading AI content immediately or add the correct label.

A proposed code of conduct on the use of AI by election observers has also been issued. According to it, observers will watch for deepfakes and synthetic content made with AI. They will inform the Election Commission immediately if they find suspicious content, and will not share it further. Observers will also not alter their observations, photographs or videos with AI tools.

The Election Commission has proposed amendments to the code of conduct prepared in 2023. Suggestions on the proposed amendments on the use of AI have been sought by 20 November.