Navigating the Promise and Perils of Artificial Intelligence in Election Management

In industrially advanced countries, former presidents attending conferences is a common occurrence, and their experience and expertise are typically valued in international and regional events.

By contrast, in Sierra Leone, President Julius Maada Bio has persistently sidelined his direct predecessor, Ernest Bai Koroma (pictured in an undated photo from the swearing-in ceremony for his second term), along with other former leaders who came to power through constitutional processes.

Recently, members of Bio's administration, some of whom have previously expressed extreme views regarding opposition supporters and parties, have advocated for the return of Ernest Bai Koroma by supporting the dismissal of charges that labeled him an enemy of the state.

These charges also placed his former appointees and family members at risk.

Koroma held the last free and fair elections in 2018, which ushered in Julius Maada Bio as president. 

Against this political backdrop, the contested 2023 presidential elections, and Sierra Leone's rankings among "narco states," Julius Maada Bio has sought to adopt artificial intelligence (AI) in election management.

"Narco countries" or "narco-states" are nations in which illicit drug cartels have deeply penetrated political, economic, and security institutions. 

Sierra Leone has articulated a vision to leverage emerging technologies to enhance transparency, voter participation, and electoral efficiency, while maintaining public trust.

This approach, presented at a major international electoral capacity-building program, reflects both ambition and caution in adapting to the evolving landscape of democratic administration.

Nevertheless, the integration of AI in electoral processes presents several risks.

AI systems may malfunction or perpetuate biases, which could undermine transparency and public trust.

Insufficient data security measures may expose sensitive electoral information to breaches or manipulation.

Additionally, excessive reliance on technology may marginalize voters who lack digital access or skills, underscoring the necessity of a balanced and inclusive approach.

During the International Capacity Building Course on Artificial Intelligence and Elections, held in New Delhi, India, an electoral officer in the Electoral Commission for Sierra Leone (ECSL) presented Sierra Leone’s vision.

ECSL participation highlighted engagement with global electoral innovation and commitment to sharing experiences and best practices.

In the address to election experts and officials worldwide, ECSL highlighted Sierra Leone’s integration of Information and Communication Technology (ICT) in electoral administration and described how AI could further advance democratic processes.

While this modernization is ambitious, it faces potential setbacks.

ICT systems are vulnerable to cyberattacks and technical failures, and resistance to technological change from institutions or the public may impede progress.

Additionally, inadequate training could reduce the effectiveness and benefits of AI adoption.

ECSL emphasized that AI tools have the potential to improve voter education and facilitation, optimize election logistics, strengthen data-driven decision-making, and enable faster communication between electoral institutions and citizens, thereby enhancing overall election management.

However, these opportunities are accompanied by significant risks.

AI algorithms may inadvertently disseminate misinformation or disinformation, and flawed data inputs could lead to suboptimal decisions.

Technical problems may also disrupt essential communication channels during critical periods.

Recognizing these risks, ECSL underscored the necessity of ensuring that AI deployment remains human-centered, transparent, inclusive, and accountable to preserve electoral integrity and credibility.

Insufficient oversight may result in unaccountable AI-driven decisions, and a lack of inclusivity could exclude certain groups.

Any lapse in transparency may rapidly undermine public confidence and the perceived legitimacy of electoral outcomes.

ECSL cited Sierra Leone’s recent electoral reforms and ongoing investment in digital technology as evidence of the country’s preparedness to adopt AI responsibly, to ensure that all voters are included.

Nevertheless, challenges remain.

Gaps in reforms or their implementation could inadvertently cause digital exclusion; persistent inequalities in technology access may continue, and limited resources could impede sustained investment in digital infrastructure.

Attendees acknowledged Sierra Leone’s commitment to innovation.

However, such recognition may introduce new challenges, including divergence between international expectations and local realities, and potential miscommunication regarding Sierra Leone’s actual readiness for large-scale AI adoption.

The ECSL emphasized that international engagement demonstrates its commitment to learning from global best practices and exploring safe AI applications to modernize election administration, strengthen institutional capacity, and improve service delivery.

However, not all global best practices are suitable for the local context, and strategies effective elsewhere may not address Sierra Leone’s specific needs.

Inadequate implementation could undermine.

The Commission reaffirmed its commitment to ensuring that technological innovation supports free, fair, peaceful, and credible elections in Sierra Leone.

Simultaneously, it acknowledged potential hazards: technological advancements may introduce new vulnerabilities, rapid progress could outpace regulatory or ethical safeguards, and public skepticism may persist if innovations are not communicated transparently or inclusively.