Wednesday / May 28 2025

Newspaper : The News

In 2024 edition of CPI, Pakistan ranked 135th out of 180 countries, scoring 27 out of 100

Corruption continues to be one of Pakistan's most pressing issues, undermining governance, slowing economic growth, and eroding public trust in state institutions.

Transparency International annually releases the Corruption Perceptions Index (CPI) as a global benchmark for perceived public sector corruption. In the 2024 edition of the CPI, Pakistan ranked 135th out of 180 countries, scoring 27 out of 100.

Several core factors contribute to Pakistan’s deteriorating position on the CPI. Foremost among these is the weakness of its institutional frameworks. Agencies tasked with ensuring accountability, such as the National Accountability Bureau (NAB), often face allegations of political bias, inefficiency and selective justice. Closely related is the pervasive lack of transparency in government operations. In the public procurement, tendering and contracting sectors, opaque procedures create fertile ground for illicit dealings.

Another critical issue is the ineffectiveness of the legal system in prosecuting corruption. Courts are often bogged down with delays, and conviction rates in corruption cases remain disappointingly low. According to data from Transparency International Pakistan’s National Corruption Perception Survey 2023, the police are seen as the most corrupt institution in the country, followed by the public procurement sector and the judiciary.

The emergence of artificial intelligence (AI) presents a new frontier in the global fight against corruption. With its capacity to analyse vast datasets, detect patterns, automate decision-making, and enhance transparency, AI offers governments and institutions a powerful set of tools to identify, predict, and prevent corrupt practices.

AI’s role in anti-corruption efforts is multifaceted. At its core, AI excels in analysing complex information and identifying anomalies that may go unnoticed by human auditors. For instance, when dealing with public procurement data, tax records or government disbursements, AI algorithms can scan millions of transactions in minutes to detect irregularities that may point to bribery, fraud, or embezzlement. Through machine learning, these systems can continuously improve, refining their ability to flag suspicious activity as they process more data.

It is pertinent to consider here how some countries are now effectively using AI to combat corruption. Brazil offers a compelling example of how AI can transform anti-corruption efforts within public procurement systems. For years, Brazil grappled with high-profile scandals, such as the infamous Operation Car Wash, which exposed deep networks of collusion between business elites and government officials. In response to these challenges, Brazilian authorities began to explore technological solutions. One such initiative was the development of Robo Laura, an AI-powered platform designed to monitor public procurement contracts. The deployment of Robo Laura enabled early detection of irregularities in public infrastructure contracts, saving millions of dollars in potential losses.

India, another country historically plagued by corruption in various sectors, has also turned to AI with notable success. In particular, the Indian government’s Project Insight is a landmark effort to enhance tax compliance using AI. India’s informal economy and widespread underreporting of income have made tax evasion a chronic problem. Project Insight integrates AI with data analytics to cross-reference information from tax filings, banking transactions, real estate records, and even social media activity. The results have been impressive: tax collections have increased, voluntary filers have grown, and human discretion has diminished.

A similarly innovative approach can be observed in South Korea, where the government has developed an AI-integrated digital platform called the Clean Portal. South Korea’s investment in digital governance has allowed it to centralise and analyse vast amounts of data across ministries. The Clean Portal uses AI to monitor financial disclosures, government contracts and civil servant activity. By comparing declared assets with procurement decisions or patterns of contract awards, the system can identify potential conflicts of interest or illicit enrichment.

One of the most striking examples of AI-powered governance can be found in Estonia, which has embraced digital transformation to a degree unmatched by most of the world. Estonia’s e-governance model integrates AI into virtually every aspect of public service delivery, from tax collection and healthcare to voting and business registration. Importantly, the country’s use of AI is not just about efficiency but also about minimising opportunities for corruption. As a result, Estonia consistently ranks among the least corrupt countries in Europe, demonstrating how digital design can align with ethical governance.

A compelling example of how AI can drive transparency in tax collection in Pakistan comes from an initiative led under my supervision by the Knowledge Economy Task Force, in collaboration with NADRA and the FBR. By leveraging FBR and NADRA’s transaction data and deploying advanced AI algorithms, the project identified 3.8 million high-income non-filers, each owing over Rs100,000 in taxes, uncovering a potential Rs1.6 trillion in unpaid income tax.

Remarkably, this three-month effort incurred zero cost to the government yet delivered powerful results: declared assets surged to Rs3 trillion, Rs65 billion in additional taxes were collected, and over 90,000 non-filers entered the formal tax system. The tax return submissions exceeded two million that year – the highest in FBR’s history. This model demonstrates the immense potential for nationwide replication, including imposing restrictions on non-filers, such as barring air and rail travel access or the renewal of passports and driver’s licenses.

Addressing the challenges associated with AI implementation is important for effectively using it. First and foremost is the issue of data quality and availability. AI systems are only as good as the data they are trained on, and in many countries, public sector data is incomplete, fragmented, or poorly digitised. Without reliable and standardised datasets, AI models may yield inaccurate results or miss key anomalies.

To use AI against corruption, our government must digitise records, ensure system interoperability and promote open data standards. Transparency is also critical; AI algorithms and decision-making processes should be accessible for scrutiny to prevent misuse or manipulation. A multi-stakeholder approach involving civil society, academia and the private sector can foster trust and innovation, while legal safeguards must be established to protect privacy and ensure ethical AI use. Building technical capacity through education, training, and international cooperation is essential to sustain AI-driven reforms.

Finally, and most importantly, curbing corruption requires political will, which is often absent for obvious reasons.