What happened
The Election Commission of Pakistan (ECP) released a draft set of 12 amendments to its 2023 election‑code of conduct that would regulate the creation, distribution, and disclosure of artificial‑intelligence‑generated material during elections. The draft, published in early October 2026, targets political parties, candidates, media organisations, and election observers. It obliges media outlets to label any AI‑created or AI‑modified election‑related content, bans deepfakes that falsely portray political leaders, judges, military personnel or other state officials, and requires political actors to disclose AI‑assisted outreach such as robocalls or voice‑cloned messages. The proposal also bars the use of AI‑powered bots or automated accounts to fabricate online engagement. The ECP has opened a public comment period that runs until 20 November 2026, inviting stakeholders to submit feedback.
In a statement released in Islamabad, the ECP outlined twelve specific changes to its 2023 code of conduct. The amendments cover four stakeholder groups: political parties and candidates, media organisations, election observers, and social‑media platforms. For media, the draft mandates a clear on‑screen or on‑page disclosure whenever AI is used to write, translate, present, or analyse election‑related material. The media would still be liable for the factual accuracy of the content, even if AI‑generated.
The draft explicitly bans "deceptive deepfakes" that portray political leaders, judicial figures, armed‑forces personnel, or other state officials in fabricated images, audio, or video when the content is false, defamatory, or intended to ridicule. Violations could be reported by PEMRA, the proposed Social Media Protection and Regulatory Authority, or the National Cyber Crime Investigation Agency, which would then forward cases to the ECP for possible sanctions.
Political parties and candidates would be required to label any AI‑generated images, videos, or audio clips used in campaign material. The draft also prohibits the use of AI‑driven bots, automated accounts, or other tools designed to manufacture artificial trends or manipulate online engagement. AI‑assisted voter outreach, such as robocalls or voice‑cloned messages, must begin with a clear disclosure that the communication is AI‑generated, and parties must retain copies of all AI‑produced material for audit purposes.
Election observers would be tasked with monitoring for AI‑related manipulation during the voting process and must report any suspected incidents to the ECP. They are also barred from using AI to alter their findings, photographs, videos, or official reports. The ECP has invited feedback from the public, political parties, media houses, civil‑society organisations, and technical experts until 20 November 2026.
Source details: minutemirror.com.pk
Why it matters
Regulating AI in elections is a growing global concern, and Pakistan’s draft represents one of the first comprehensive attempts in South Asia to embed AI‑specific transparency and anti‑misinformation safeguards into electoral law. If adopted, the rules would create enforceable obligations for media houses, political campaigns, and election monitors, potentially reducing the spread of deepfakes and automated propaganda that can distort voter perception. The involvement of multiple regulatory bodies—including PEMRA, the proposed Social Media Protection and Regulatory Authority, and the National Cyber Crime Investigation Agency—signals a coordinated approach to monitoring digital content. However, the draft leaves several practical questions unanswered, such as the penalties for non‑compliance, the technical capacity of the monitoring agencies, and how the rules will be applied to cross‑border platforms that host AI‑generated material. The public‑comment window also offers a rare opportunity for civil‑society groups, technologists, and international observers to shape the final policy.
The amendments aim to close a regulatory gap that has allowed AI‑generated misinformation to spread unchecked in previous Pakistani elections. By mandating disclosure, the rules could increase voter trust in media and campaign communications, reducing the impact of fabricated content that might sway public opinion.
The coordinated oversight involving PEMRA, a new social‑media regulator, and the cyber‑crime agency reflects an acknowledgement that AI‑related threats cross traditional media boundaries and require a multi‑agency response. This model could serve as a reference for other emerging democracies grappling with similar challenges.
Nevertheless, the draft does not specify the scale of penalties, the resources allocated for detection, or the legal standards for proving intent behind deepfakes. These ambiguities could hinder effective enforcement and may invite legal challenges on grounds of free speech or over‑broad regulation.
The public‑comment period provides a window for experts to propose technical standards—such as AI‑generated media or establishing verification protocols—that could shape the final rulebook. The outcome will influence how evolves in a region where digital literacy and internet penetration are rapidly expanding.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
A route planner searches possible journeys using explicit rules. What does this illustrate about AI?
What to watch next
Key indicators to follow include: (1) the final version of the amendments after the November comment period, especially any changes to enforcement mechanisms or penalties; (2) how PEMRA and the new Social Media Protection and Regulatory Authority develop technical guidelines for detecting undisclosed AI content; (3) reactions from major Pakistani media outlets and political parties, which will reveal the feasibility of compliance; and (4) any legal challenges or court rulings that may arise if the rules are perceived to infringe on free‑speech protections.
The final amendment text after the November comment deadline, especially any revisions to enforcement provisions.
Guidelines or toolkits released by PEMRA and the Social Media Protection and Regulatory Authority for detecting undisclosed AI content.
Statements from major Pakistani news outlets and political parties indicating readiness or resistance to comply with the new disclosure requirements.
Potential legal challenges filed in Pakistani courts questioning the constitutionality or scope of the AI‑related provisions.