corruption in pakistan

Corruption starts when a government employee deviates from the law and favors someone out of the way. It could be either for the financial benefit or the power or sometimes both.

Unfortunately, in Pakistan, this practice has engulfed the entire governmental procedures to the extent that sometimes, even the cleanest practices are also considered shady due to the extra load corrupt practices have put on every single employee.

Then how to start to end this?

There is a way around, as the technology is becoming smarter and Artificial Intelligence (AI) is playing role effectively across the globe, Pakistan can take advantage to streamline everything once for all.

All over the world, AI has proven to be more efficient in computing large pools of data. Giants like Google and Facebook use AI to organize and maintain big data so why not us? Why not use cloud computational techniques and big data processing to end this menace of corruption?

In Jan 2018, researchers from the University of Valladolid (Spain) created an Artificial Intelligent (AI) system that can predict the areas that have a higher risk of corruption. A similar system can be established in Pakistan too.

We can evolve to make a similar system by just starting with the following;

Create and Maintain Organized Database

1- Payroll data of all government employees currently stored in “Project for Improvement to Financial Reporting & Auditing” (PIFRA) System should be complete, correct and CNIC based as opposed to PIFRA account number.

2- There shouldn’t be more than one PIFRA account number assigned against a unique CNIC. Income other than the salary should explicitly be mentioned to keep the record of all known sources of income of every govt. employee.

3- Financial records of all the govt. employees, including bank account details, as well as the details of dependants, should clearly be listed in the database. CNIC and/or CRC (B-Form) based data of all dependents of every Government employee should be maintained, after cross verification from NADRA.

4- Complete Service history and data regarding transfers, postings and take home salary should also be included.

How This Data Will Help To Detect Corruption?

Digitizing such information in a database will help better organize data, and with the help of preset protocols, AI will be able to flag where the corruption occurs in the system.

1- With a large pool of data from various platforms, AI will be able to cross check the already registered data to every govt employee. For instance, Data from Land Revenue Management Information System (LRMIS) could be used to check the sale and/or purchase of property against CNIC of Government Employee or any of his/her dependents.

2- Other than this, Artificial Intelligence will also help keep a smooth flow in maintaining financial data by having access to State Bank’s Financial Monitoring Unit (FMU) and checking account balance of all bank accounts of every Government Employee and his/her dependents to ensure transparency.

3- Moreover, Data from Tenants Registration System could be pulled every month to check if any Government Employee or his/her dependent is showing up as the owner of a rented property.

4- Similarly, data from excise and taxation will help detect whether any govt employee or his/her dependent is showing up as owner of how many vehicles and if it exceeds the payroll of the employee, AI induced system will mark corruption instantly.

5- It would also be easy to keep a check on abroad travels by any govt. employees or his/her dependents by pulling out data from FIA.

With systems talking to systems, a report that identifies individuals presumed to be living beyond declared sources of income is easy to generate. But it is equally important that this aggregated data is not abused by individuals who have access to it. In this regard, it is proposed that Artificially Intelligent Bots may be used to generate this encrypted data-anomaly report every month and send it to concerned individuals, with the highest level of clearance, who could decrypt the data.

On the other hand, there should be a timeframe within which these individuals can digitally correct previously submitted information and/or add new Information. After the expiry of the allotted time, those still on the data-anomaly radar may be called in to explain the reasons for the reported anomaly, before deciding their fate.

image credit:

The author is the member of Prime Minister's Task Force on Austerity & Structural Reforms

Leave a Reply

5 Comment threads
0 Thread replies
Most reacted comment
Hottest comment thread
5 Comment authors
Samana AliFahd ChishtyNadia F. AlviWaqas A. KhanNaeem Malik Recent comment authors
newest oldest most voted
Notify of
Samana Ali
Samana Ali

Undoubtedly there is an explosion of real-world applications of artificial intelligence and machine learning that are changing the game of governance in today’s world but Pakistan is among a long list of developing countries that are not given necessary access to their countries’ own data. The view on corruption is quite limited in the article above where we are assuming that the government officials will use a) legal means to hide their corruption b) there is increasing automation, accuracy, and frequency across technological processes in the government sector. We have to start the conversation with the need for developing new methods for data development that are best suited to the current technological standing of our country. There is a need to broaden our understanding of big data and how we can develop new data management techniques to prevent fraud and abuse in the public sector (much more complex than cross-checking the accounts of the public servants and their “registered” assets). The introduction of data mining is also required to check one of the biggest areas of corruption, public procurement. It can help in auditing practices to monitor when the government or public servants are issuing bids and to identify red… Read more »

Fahd Chishty
Fahd Chishty

For the system above mentioned it is imperative that existing databases be employed optimally for reduction in time of effectiveness.
Systems should be able to talk to each other on a basic level
Block chain capability can be employed for integrity, confidentiality and availability of the system.

Nadia F. Alvi
Nadia F. Alvi

It is an interesting approach that is achievable as well. However, with the data inflow from multiple departments, it might require a certain level of data re-engineering to maintain consistency but opting for a centralised information system would be more practical to avoid subsequent updates. Cloud computing (hybrid if not cloud) would a great way to tackle data sensitivity and loss (that we experience quite often in Pakistan). One thing to look for is to identify accounts (financial transactions) not on the name of employee/dependents. Overall, it is a practical approach that you have suggested and with a robust implementation, I am sure you would do justice as you did with other ventures. Good luck!

Waqas A. Khan
Waqas A. Khan

Dear Burhan Rasool, While I am convinced that you are an excellent asset of Pakistan and have done wonders at PITB, your this article is much novel as well. Interesting in fact. However let me add my input into this. 1. Your suggestion about connecting PIFRA to CNIC is excellent and then a subsequent suggestion to make it unique as well is superb. However, when you talked about tracking income from all knows sources, i felt that this needs more elaboration. For me, tracking expenditure is easier than tracking income because keeping in view the currect practices in Pakistan, it will be much harder to track money from unknown sources as no one govt. Employee will do this volunterily. So for this, in this article you can add one more suggestion and that is converting the cash system to plastic money as being used in developed countries. Instead of trying to track how they earn which the majority wont tell as FBR is still failing to check the correct income of every Pakistani because they lie with confidence, lets make the system as such that from a child to adult, 90% of the transactions shall come through banking channels, credit… Read more »

Naeem Malik
Naeem Malik

Dear Burhan, First it needs to identify different levels and patterns of corruption. You discussed only one aspect based on personal data of an employee. But I think most of the corruption involves in procurement. That money may be sidelined easily without letting it appear on the personal radar. PIFRA, State Bank, FBR etc are already doing most of the job you suggested here. Please come up with broader spectrum of the issue. Regards and best wishes.