Go to content
Featured
NICTScience

Gayo Diallo, the Guinean researcher who wants to make voting more reliable thanks to artificial intelligence

Gayo Diallo, teacher-researcher at the University of Bordeaux and expert in information systems © Gayo DIALLO

In many low- and middle-income countries, electoral processes remain marred by dysfunctions that undermine public trust and democratic principles. Fraud, lack of transparency and post-election violence are becoming recurrent. To remedy this, civil society is trying to act, but its means often remain limited.

It is in this context that Horoyah, an innovative platform designed by the Guinean researcher Gayo Diallo, teacher-researcher at the University of Bordeaux and expert in information systems.

A citizen platform to monitor elections

Horoyah is based on the use of knowledge graphs, a technology resulting from symbolic artificial intelligence which makes it possible to structure, link and analyze large quantities of data. “A knowledge graph helps organize data into nodes and links between nodes, which represent simple facts,” explains the researcher. “Its particularity is to integrate formal semantics, that is to say a meaning understandable by machines allowing reasoning and inferring new knowledge. »

In other words, this technology is not limited to storing information, it understands it and links it logically, a bit like Knowledge Graph from Google, which allows the search engine to directly answer questions such as: What is the population of Guinea?

Thanks to this innovation, results reported by citizens from polling stations can be collected, verified and visualized on interactive maps. Horoyah thus facilitates the identification of irregularities, the detection of trends and the creation of an auditable and transparent database.

The case of the 2020 presidential election in Guinea

To illustrate its effectiveness, the researcher applied Horoyah to the 2020 Guinean presidential election. While the Independent National Electoral Commission (CENI) proclaimed the victory of Alpha Condé with 59.49% of the votes, the data collected by civil society presented notable discrepancies.
Horoyah made it possible to highlight these differences and of raise the question of transparency in the aggregation of official results.

According to Gayo Diallo, “with this framework, it is easy to measure, for example, the difference between the sum of votes declared in a region and that resulting from the aggregation of field data. »
The system can also cross-reference election results with other sources, such as population density, to “to identify contradictions or potential fraud”.

Guarantees against false declarations

Faced with the risk of misinformation, the platform integrates several safeguards.
“Each citizen declaration must be accompanied by proof, in particular the image of the minutes of the polling station,” specifies the researcher.
Before being recorded in the system, this information goes through a validation step : “The reported data is first verified in an environment SaaS before their final registration. »

This rigor ensures the reliability of the information and strengthens the credibility of the process.

Adapt technology to local realities

One of the major challenges remains limited Internet access in rural areas.
But for Gayo Diallo, solutions exist: “We can easily imagine data transmission by SMS via 2G or the UDDI protocol, which does not require an Internet connection. In this case, the sending of images would not be instantaneous, but the textual results could still be integrated. »

Train users to ensure efficiency

Horoyah requires different levels of training depending on user profiles.
“Those who are in the field need minimal training: knowing how to fill out a form and take a photo,” explains the researcher.
“On the server side, training focuses on system validation and maintenance. Finally, updating and evolving the graph requires a more technical level of expertise. » This graduated approach makes it possible to involve citizen observers as well as engineers and data analysts.

Towards the future of digital elections

Horoyah opens up promising prospects for strengthen electoral transparency and citizen trust, but there is no shortage of challenges: disinformation, cybersecurity, digital divide and even data reliability.

For Gayo Diallo, the key lies in local adaptation and technological evolution:
“The platform already relies on artificial intelligence, but it could go further. Ultimately, it could integrate modules capable of automatically detecting anomalies or even allowing citizens with low literacy skills to participate via voice tools. »

Such an innovation, according to him, could transform the way elections are monitored in Africa and elsewhere, by giving more power to citizens and civil society organizations.

You can read the article in the scientific journal here: https://zenodo.org/records/17444454

Alpha Oumar Bagou Barry

Share

Read also

Leave a Reply

Your email address will not be published. Required fields are marked *