Country Profile – Pakistan

With the introduction of a new strategic plan for 2025–2029, the Election Commission of Pakistan (ECP) is markedly expanding its use of AI across operations. While the number of AI systems in full deployment is still limited as of early 2026, key initiatives that will impact core voting arrangements are actively being designed, developed, procured and piloted.

The ECP’s exploratory approach to AI casts a wide net across various existing processes, to identify where AI can improve them as well as existing software. This approach is notably effective, as the ECP has steadily digitized its operations since the early 2000s. In recent years, it has introduced proprietary systems such as its election management system (EMS) for result tabulation and transmission and a computerized electoral roll system (CERS) for the automation of voter registration and verification (ECP 2024; Election Commission of Gilgit-Baltistan n.d.).

How is AI used to improve electoral management in Pakistan?

While the Election Commission of Pakistan is undergoing a transformative reconfiguration of how AI is used across the Commission in the short to medium term, there are already a select number of ongoing AI-related use-cases. Particularly, the recent cross-agency deployment of AI facial recognition in voter roll management marks a significant step towards integrating AI to improve the integrity of the ECP’s operations.

At present, AI is deployed in two principal ways:

•  Voter roll management: Since 2012, registration officers at the district level have used computerized electoral rolls systems (CERS) to link voter roll data to a centralized server, ensuring greater coherence and integrity in the system (International Foundation for Electoral Systems 2013). Now, the Commission are developing AI-based systems to identify voters using family tree data. These tools will verify whether an individual legitimately belongs to a particular family, preventing the inclusion of unrelated persons in family records. As a result, voter lists will be organized so that members of the same family are correctly grouped together.

• As productivity tools: Across the ECP, staff members are already applying AI-powered productivity tools to assist them in everyday operations and routine tasks. These include note-taking, summarizing and analysing text, editing, and internal communications.

   

   

Areas where AI tools are currently under consideration

In 2025, the ECP introduced a new strategic plan for 2025–2029, involving a range of initiatives that incorporate AI-powered technologies. While ongoing projects are at different stages of conceptual development, they jointly illustrate a significant commitment by the ECP to investing in AI technologies for a wide range of operational areas in the coming years.

Key examples of cases where the Commission is looking to integrate AI include:

• Voter list management: The ECP is investigating how machine-learning technologies may help connect voters to their nearest polling station by matching them with the correct census code. Historically, voter list data have been collected manually, door-to-door across the country’s 180,050 census block units. The process has been difficult to streamline, since each individual voter needs to be matched with a corresponding census block code that assigns them to their most convenient polling station. By integrating automated systems with rule-based algorithmic instructions, the ECP is expecting to reduce non-systematic mistakes significantly, thereby improving the reliability of systems.
• Constituency delimitation: In conjunction with the collection of voter data, the ECP also manually delimits the constituencies of general and local elections based on a balanced distribution of voter numbers across geographies. The delimitation of constituencies is often subject to public scrutiny. Upon publicizing the final results of a delimitation exercise in 2022, the ECP reported receiving nearly a thousand objections or representations (Free and Fair Elections Network 2024). By automating the delimitation using AI, the ECP aims to ensure that the process aligns in an objective way with the principles of proportionate allocation as stipulated in Pakistan’s Election Act (Pakistan 2017).
• Social media monitoring: In order to address the escalating challenges of electoral mis- and disinformation on social media platforms, the ECP intends to implement commercial social monitoring and sentiment analysis tools, allowing it to address misleading narratives before they gain traction. Specifically, these platforms will help the ECP stop the spread of misinformation relating to election logistics—such as erroneous dates, or falsehoods about polling station locations and availability—that risk disenfranchising voters or undermining public trust in the democratic process. The acquisition of the AI tool is currently under review, to ensure that the third-party software complies with the ECP’s technical and regulatory requirements.
• Call centre and chatbot: Currently, the ECP has a call centre that utilizes an interactive voice response (IVR) system for navigating options and rerouting callers. To expand the utility, convenience and accessibility of election information services, the ECP may consider options to introduce AI capabilities to call-centre phone lines, as well as incorporating a large language model (LLM) chatbot into the ECP’s website for the simpler type of questions.
• AI-generated training materials and handbooks: The ECP is exploring the use of LLMs in generating handbooks or other training resources for staff—specifically for polling stations staff. Although AI systems may contribute additional interactivity or pedagogical elements to the materials, the content of the training programmes is always rooted in existing internal policies.
   

Building institutional AI literacy: training the trainers

The ECP is adopting a cascading approach to the introduction of AI systems throughout the Commission. Initially, there is investment in specialized technical staff, who oversee the early stages of conceptualizing, assessing and evaluating potential operations in which AI may be integrated. At this stage, the Commission is looking to expand AI expertise in its IT and innovation wings by conducting educational courses that improve staff capacity to train, design and develop models in-house. Ultimately, these training programmes aim to foster technological autonomy, so that the ECP can pursue AI-related initiatives with a greater degree of institutional independence from external service providers. 

Once a solid foundation of expertise within the ECP’s technical divisions is secured, specialized IT staff will, in turn, ‘train the trainers’, i.e. train technical staff at different levels across the Commission. In the long term, technical staff will then disseminate their knowledge of AI to remaining non-technical staff members, particularly those who work closely with AI-powered systems. This training strategy is formally integrated within the Commission’s strategic plan, which identifies raising AI literacy as a time-limited target subject to reporting within the coming five-year period.
 

AI policy frameworks and the ECP

The ECP is involved in several parallel processes independently, in order to develop AI policy frameworks and align them with various externally mandated governance frameworks relating to AI in elections.

• National AI Strategy: While the ECP’s strategic plan for 2025–2029 informs the practical implementation of AI systems on a needs basis, the Commission is also subject to Pakistan’s National Artificial Intelligence Strategy, which was published in 2025. This instrument sets out measures to raise public awareness and readiness among public servants in terms of AI literacy, ethics, data governance and personal data protection. Additionally, it specifies efforts to ensure human oversight, regular auditing, public transparency and end-to-end cybersecurity protection for high-risk processes, including elections (Ministry of Information Technology and Telecommunication 2025).
• Code of Conduct and legal amendments: In response to the growing use of AI in political campaigning, and its potential to amplify mis- and disinformation, the ECP is actively revising the Code of Conduct for National Media, which was released initially in 2023 (ECP 2023). These updates specifically aim to address ways in which AI may exacerbate the magnitude of information pollution in the electoral environment. Alongside these revisions, the ECP has submitted proposals to Federal Government for legal amendments that would require mandatory information, including about provenance, in the labelling of  AI-generated materials, helping to reduce the influence of deceptive content.
• Public awareness through policy: On top of outlining benchmarks and targets for the practical implementation of AI, the strategic plan for 2025–2029 also emphasizes that the ECP’s responsible governance of electoral AI involves raising public awareness. Specifically, this entails raising awareness about the impact that AI has on elections, in terms of how it is integrated into electoral administration, as well as how it is used by other actors in the electoral sphere for political campaigning or influence. By increasing AI literacy among the public, the ECP would ensure that voters grasp how the adoption of AI impacts all aspects of the electoral process, from electoral administration to algorithmic information feeds and AI-generated content.