Speech and Elections

Our panelists were grateful for the active Q&A session but ran out of time before they could address all of the questions and issues raised by attendees. They have responded to a select number here. Some questions have been lightly edited for clarity and context.

Question: While cyber laws and proposals around labelling may assist in combating fake news, these may not offer immediate tools for an EMB monitoring campaigns in high stakes electoral processes. In your experience what has been the most reliable strategy for EMBs? Especially where the EMB has direct mandate to monitor and regulate political party campaigns and are faced with possible fake news-false information-(bordering on violations of the Elections Code of Conduct) that directly impact the EMB's mandate?

Answer: The most reliable strategy is usually a tiered response that combines providing authoritative information and narrow, mandate-based enforcement. In practice, this means the EMB should draw a narrow line on what types of information that it acts on – and specifically anything that directly affects its mandate. This includes information about voting procedures, candidate registration, results, or clear campaign code violations. The other key consideration is to document criteria and process. Where EMBs do have a regulatory role, it’s important to emphasize transparency: what standards were applied, what evidence was used, who made the decision, and what review is available. 

Question: How can we use AI to enhance voter education, especially among voters with less education?

Answer: AI can reduce the costs and time needed to translate educational content into different languages and reach members of minority groups. It can also be used to generate supporting images and audio that increase their accessibility to illiterate people or persons with disabilities.

During our presentation, we talked about AI being used to scrape private information on media users (such as psychological profiles, political leanings, and consumer habits) for political persuasion purposes but, if done ethically, AI-powered data collection can also help election commissions more easily identify knowledge and information gaps and tailor messaging accordingly.

Question: In June 2025, IFES held an event in Washington, D.C., called the “Launch of the AI ​​Election Advisory Group (AI AGE).” Can you provide an update on this group, and where can we learn about its work?

Answer: IFES' AI AGE is drafting a Strategic Playbook on AI and Elections, to be released in the coming months. AI AGE members also participated in the first stakeholder consultation with the UN AI Global Dialogue Co-Chairs earlier this year, arguing that AI and elections should be elevated as a workstream under the upcoming AI Global Dialogue in Geneva. IFES will announce some more updates on this dynamic body of work in the coming weeks.

Question: Given the depth of these reports, what specific entry points do you see for democracy-focused civil society organisations to move beyond just monitoring AI-driven disinformation, start influencing the regulatory frameworks that govern political communication and hold political actors accountable?

Answer: CSOs can engage earlier in electoral and digital governance reforms, especially where rules on political advertising, synthetic media, and transparency are still being defined. CSOs can also influence “soft law,” the quasi-legal instruments such as election commission guidelines, party codes of conduct, and platform policies, where practical standards are often set first. During elections, CSOs can act as an accountability bridge by documenting patterns of misuse or selective enforcement.

In addition, CSOs can translate monitoring findings into structured evidence for regulators and post-election reviews, helping identify systemic gaps. They can also improve accountability by publishing public report cards, encouraging parties to make commitments, and using complaint mechanisms, while working together with technical experts and media to help improve rules on political communication. 

Lastly, a key entry point is leading on voter resilience efforts, like digital literacy and trusted information sources. Ultimately, the impact of CSOs comes from working across policy, practice, and public engagement to ensure that responses to disinformation and AI strengthen democratic participation.

Question: How can AI be used during elections to minimize violence and disputes during political campaigns?

Answer: We would recommend looking at the root causes of electoral violence first. GenAI could maximize disinformation (think of AI-generated videos showing fabricated electoral fraud), which could instigate violence. 

But AI can also help election officials and other stakeholders reduce the incentives for engaging in violence or addressing conflict. This could include more effective early warning systems and the ability to process large amounts of data to inform the allocation of security personnel on election day. It could also include conducting sentiment analysis to assess and calibrate public discourse, and enabling citizens and political actors to file complaints more effectively. The use of AI could also help courts triage and process these complaints faster, potentially addressing grievances through formal means. 

Question: From an African perspective, where regulatory capacity and digital literacy gaps still exist among citizens, I am curious: what practical low cost strategies can  electoral bodies adopt to both detect AI-driven misinformation and build public trust without overlying on restrictive information laws that may unintentionally limit civic space? 

Answer: If EMBs have dedicated personnel, they can, for instance, engage in “social listening” or set up lightweight monitoring and escalation workflows to detect misleading or harmful narratives gaining traction online. This in turn can then inform their authoritative, evidence-based narratives instead. 

For instance, if people seem to be engaging with online posts that undermine trust in a certain part of the process – say, out-of-country voting – the EMB could invest more time and resources into explaining security features embedded in the OCV process. This effort can focus on educating voters without necessarily touching on the disinformation piece and giving it more exposure. 

Another important strategy: it is key to invest in partnerships to amplify accurate information and build credibility. EMBs can also work with civil society organizations, academia, and other reputable organizations that already do this social media monitoring and fact-checking, getting insights on misleading narratives from them and also providing them with factual information to help counter falsehoods. People who do not necessarily trust the EMB or public institutions might be more open to hearing the technical opinions of those trusted institutions in their communities.