Problem description
In March 2013, Kenya, a small country in East Africa, held its general election to elect a new president. There were two main candidates vying for the post; Uhuru Kenyatta and Raila Odinga. Prior to the election process, the country had installed a digital system, which would help make the voting and tallying processes seamless (Curtis 1).
The system had been touted to reduce the time taken to count the votes. In this system, the polling officers would key in the data from the ballot paper count into a system that would send the information to one national tallying center (Marks 1). The polling officers would then call the central tallying center and present the same information by voice, for confirmation purposes.
The system worked well for the first few hours after the polling stations closed. The country was keeping track of the tallying process because the data was updated in real-time and broadcast on live television (Marks 1). However, as the data kept increasing, the system started getting overwhelmed (Warigia and Longwe 1).
The digital portal that had been installed to collect information from remote polling stations collapsed. The system was at one point reported having been multiplying the actual vote count by five (Curtis 1). The mobile phone network that had been dedicated to communication between the electoral commission’s officers also got jammed because of the high number of phone calls and short messages it was forced to handle.
Given the fact that the country had gone into the election on the back of political tension, any lapse in the system was viewed as an attempt by one of the competing candidates to rig the election (Marks 1). The government officials running the systems had been believed to be partisan, and every time the system failed, social media was awash with complaints by supporters of both candidates.
Supporters of Uhuru said Raila had sent hackers to weaken the system and feed data that would give him an edge. On the other hand, Raila’s supporters accused the government of having intentionally damaged the system in order to modify the data in favor of Uhuru (Warigia and Longwe 1).
After 36 hours of counting, the electoral commission’s chief executive officer announced that they were re-starting the tally because of the failure of the digital system (Curtis 1). The new tally was to be done by getting all data from the remote centers to one point in the country’s capital Nairobi and having it manually totaled. Various parties went public to complain about the ensuing disorganization because the digital system had been made to appear secure (Warigia and Longwe 1).
After the manual count, Uhuru Kenyatta was declared the winner and was subsequently sworn in as president. His rival, Raila Odinga, went to court to contest the outcome of the election, arguing that the damage in the digital system was intentionally caused to give his rival an advantage. After a month-long case, the court concluded that Uhuru had won fairly. But the country was left divided as to the result, given the fact that the system collapsed as the country watched on television.
In this case, the problem arose as a result of failing communication between humans as a result of over-reliance on information and technology. The electoral commission in Kenya had set up the new system and, even without testing it, placed all their hope on it to make their work easier (Marks 1). These officials did not foresee the possibility of damage in the system and failed to install proper measures to deal with any reports of failure.
By broadcasting the outcome of the tallying process as data streamed in, the commission proved that they had full confidence in the system. Unfortunately, for them, the system broke down as the public watched, and the officials had to suffer the embarrassment of informing the nation that their much-publicized systems had collapsed. Even the international community was left questioning why the commission had to place all their hope in one channel (Marks 1).
Numbers to confirm the problem
In July 2013, two months after the polls, Kenyan research firm Infotrak conducted a study to find out how Kenyans viewed the country’s electoral commission in view of the year’s election outcome (“Infotrak: Most Kenyans”). The data collected showed that 53% of Kenyans did not think the election was free and fair, attributing their opinion to the inefficiency by the Independent Elections and Boundaries Commission (IEBC) in managing the polls (“Infotrak: Most Kenyans”).
The report showed that the system was operating at less than 30% efficiency. As a result, 28% of the country’s voters no longer trust the IEBC (“Infotrak: Most Kenyans”). Kenya has 14 million registered voters. If 28% of these people do not have faith in the electoral system, as shown in the Infotrak report, this translates to 3.8 million people.
During the last elections, the winner had about 5 million votes. With 3.8 million voters expressing their dissatisfaction with the process, if elections are held today, the winning candidate will certainly take home less than 3 million votes, making him an unpopular president (“Infotrak: Most Kenyans”). Fixing the communication problem could help tame this ripple effect.
Causes of the problem
The collapse of the networks in the Kenyan election situation described above was primarily a result of human error. The people tasked with feeding information into the system had not been properly trained and ended up keying in the data more than once. These multiple entries overloaded the database that had been set up to handle data from the tallying centers.
In addition, the mobile phone network, which had been dedicated to the exercise, ended up getting jammed because no order of call had been established. Returning officers from various parts of the country were calling and sending short messages to the main tallying center at the same time, leading to its early downfall.
Another problem, which was witnessed in the election vote count, came in the form of the live television feeds. Giving the public access to raw data from the tallying centers made it difficult for the commission’s officers to provide a credible explanation whenever anything went wrong. A case in point is when the system developed an error that made it multiply the total tally by a certain percentage. The wrong data was immediately noticed by members of the public, a section of whom was getting agitated by the fluctuating numbers.
The country had witnessed extreme post-election violence in 2007, and in 2013, the world was watching to see whether the situation could be averted by ensuring smooth voting and tallying. With the numbers being unstable, there was a great possibility that any errors on the part of the regulating organ could lead to an eruption of anger and consequent violence.
Possible solutions
To prevent a recurrence of the crisis that arose in the situation described above, a number of solutions need to be effected before the next election is held. First, is the training of the people manning the systems. Information, communication, and technology (ICT) systems are only as good as the people running them. The polling officials need to be trained properly on how to run the systems and what to do in case they make an error.
The system also needs to be developed in such a way that it does not allow multiple entries of the same information. Technicians need to be placed at every input center to help correct wrong entries. Getting trainers from the countries where the systems were made to conduct the training sessions would ensure that the handlers of the data are well equipped to deal with any challenges that may come their way.
Because the data keyed in through the system was verified using phone calls, a network that can handle heavy traffic needs to be set up and tested well in advance. The company providing the service needs to put the system through the toughest challenges possible and develop quick recovery solutions in case of a failure. The strength and ability of this system should be verified by independent evaluators, preferably from other countries.
Connecting the data feed to live television networks was also a big mistake, which should not be repeated. All data should first be ascertained at the tallying center before being sent out to broadcasters. This way, when wrong data is received, or when the tallying system suffers a breakdown, it can be fixed without alarming the public.
For instance, when the system started multiplying incoming data by a percentage, the bug could easily have been rectified without the public’s knowledge and then the proper tally given to television networks. It was embarrassing for the electoral commission to have the mistake pointed out by viewers, hours after it had occurred.
These technological solutions can only be confirmed effectively after being taken through various tests and subjected to the harshest of situations. Any communication problems, which are detected during the test phase, can be rectified until the system is confirmed to be working perfectly.
Even with all these systems in place, the best solution for this communication problem would be to provide easy ways of delivering the forms containing the numbers from the remote centers to the main tallying center in Nairobi.
By ensuring that polling officials can physically present the data at the national tallying center, chances of the data being intercepted and manipulated by hackers are reduced. In this regard, the government needs to dedicate light aircraft and helicopters to ferry in data from far-flung regions and fast vehicles to deliver electoral officials from nearby stations.
Selection of solution
In Kenya, election time is usually an emotion-filled period, and any small mistake on the part of the officials can lead to chaos. It is, therefore, imperative that a fitting solution is identified to rectify the problem that was witnessed in March 2013. This report has shown that the situation that occurred in the 2013 Kenyan election was as a result of a miscommunication between officials stationed at different polling centers around the country and those in the national tallying center, due to a collapse of the installed ICT systems.
The Kenyan government needs to realize that while it is desirable to rely on technological advancements to speed up processes, it needs to invest in programs to test out the systems before deployment. With the next election taking place in three years’ time, the government should make sure that it sets up the necessary networks, which should then be tested over the next two years for adaptability to different load capacities.
In the case of mobile phone networks, the government should team up with the major telecommunications firms in the country to develop a system that can handle heavy traffic without collapsing. This is because these companies (Safaricom, Airtel, and Orange) already have the expertise and resources needed to develop a strong system. The electoral commission’s officials should also put structures in place to ensure that the outcome of the polls is presented at the main office on time. This includes setting aside a budget to cater for the air freight.
How the solution could be implemented
In order to set up a properly working digital voting system, IEBC needs first to get rid of all the gadgets it bought for the 2013 election and place bids for the supply of new systems. The company with the most experience in setting up networks for elections in developing countries should be selected. This is because the challenges Kenya experienced in 2013 must have been reported in other countries while deploying similar systems.
Once the right company has been identified, it should deliver the necessary tools and equipment to the country and set up the systems as required. Individuals should be recruited and properly trained on how to run the system. Possible scenarios of failure should be created during the training process, and the handlers educated on how to deal with them. After the system has been deployed and the people to man it recruited and trained, a primary test should be conducted.
IEBC should recruit a group of people from across the country to feed data into the system. Simultaneous input of data from various parts of the country should be used as a way of testing the stability of the system. If the system remains stable even during times of heavy data input, it should be recommended for use in upcoming elections. However, if the system creates a gap in communication as it did in 2013, it should be evaluated for loopholes and any breaks fixed, before a second try.
The next item in the preparation phase is the setting up of a strong and properly functioning telecommunication network. By speaking to the mobile service companies in the country, the Kenyan government will decide on whether or not a new network that runs parallel to those being run by the main players needs to be set up. If the experts recommend the establishment of a new network, it should be set up immediately and taken through trial runs to test for efficiency.
However, it may be cost-effective for this new network to be hosted on some of the already available resources. The telecommunications network needs to be tested by recruiting teams of individuals to make calls or send short messages simultaneously. This exercise will help determine whether or not it can handle the bulk of the load during election time.
If the network allows for proper transmission of information, it should be recommended for use in future polling exercises. On the other hand, if it fails, engineers should go through it and identify the weaknesses, which should then be fixed before more trials are conducted to confirm its stability.
Works Cited
Curtis, Ben. “Kenyan voters frustrated with delayed results after electronic system fails.” The Guardian. 2013. Web.
“Infotrak: Most Kenyans lack confidence in IEBC and Supreme Court.” Evening news. Kenya Television Network, Kenya, 2013. Web.
Marks, Joseph. “How technology failed to fix Kenya election.” Next Gov. 2014. Web.
Warigia, Bowman and B. Longwe. “Technology, transparency, and the Kenyan general election of 2013.” Al-Jazeera. 2013. Web.