Cyprus seeks EU action to block AI interference in 2028 elections
Cyprus and seven other EU states are pushing for stricter oversight to prevent deepfakes and malicious AI from manipulating voters.
The short version
- Cyprus and seven other EU states issued a 'non-paper' calling for stricter AI oversight and protection against impersonation.
- The 2028 election faces risks from deepfakes, anonymous paid advertising, and chatbots offering voting advice.
- Cyprus has a high rate of generative AI use, with 44.2% of people using such tools in early 2025.
- The country currently lacks a dedicated institution to respond to hybrid attacks during active election campaigns.
Why it matters
As AI tools become easier to use, the risk of 'hybrid attacks'—where fake content is used to undermine trust in democratic results—increases. For voters, this means the integrity of an election could be compromised by content that is impossible to debunk in real-time.
A growing threat to democracy
Cyprus is preparing for the 2028 presidential election in an environment that differs significantly from past cycles. The country faces a landscape where fake audio, fabricated videos, and anonymous networks of paid advertising can be produced in mere minutes. These tools allow malicious actors to influence voters through chatbots that offer specific voting advice or by spreading coordinated disinformation.
The threat is not theoretical, as several European nations have already experienced the consequences of such manipulation. In Slovakia, a fake audio recording circulated during a pre-election silence period, while Romania had to annul its first presidential round due to coordinated TikTok manipulation. Ireland also saw a deepfake news bulletin falsely claim a candidate was withdrawing from the race.
Cyprus experienced its first major taste of this threat in January when an eight-minute video reached thousands of viewers almost instantly. It took an official body six months to conclude the video was a hybrid attack. This delay highlights a critical lack of mechanisms to warn voters quickly if a similar attack occurs immediately before a vote.
The push for EU oversight
Cyprus has joined seven other EU member states in calling on Brussels to take decisive action against the misuse of artificial intelligence. The group circulated an informal policy paper, known as a non-paper, to advocate for stricter oversight and greater transparency in platform algorithms. They are specifically seeking measures to protect against impersonation and the malicious use of AI technology.
President Nikos Christodoulides acknowledged these dangers during a summit in Dublin, noting that malicious actors are exploiting technology to influence results across the bloc. He warned that the region is vulnerable to these technological threats, especially as 2027 will be a major election year for many European countries. The summit also addressed concerns about systemic consequences for critical infrastructure from cyberattacks.
The primary goal of the eight countries is to move beyond simply analyzing attacks after they happen. Christodoulides argued that governments have an obligation to have a framework in place before the damage is done. The focus is on ensuring that authorities have the tools and credibility to tell voters what is real within hours of a deepfake appearing.
Vulnerabilities in the system
Cyprus currently lacks a dedicated body responsible for responding to hybrid attacks during an active election campaign. While the country has a Digital Security Authority for general cybersecurity, there is no specific institution tasked with monitoring election-specific disinformation. The current system relies on the Interior Ministry to organize elections and the Cyprus Radio Television Authority to oversee online platforms.
The country's technological landscape makes it particularly susceptible to these types of interventions. According to Philenews, Cyprus ranks seventh in the EU for generative AI use, with 44.2% of people using these tools in early 2025. This is significantly higher than the EU average of 32.7%, meaning a large portion of the electorate may rely on AI-generated content for information.
The 2028 election is expected to be a particularly difficult test because it involves individual candidates who are easier to target. With an electorate of roughly half a million, even a few thousand votes could decide the outcome. The week between the first and second rounds is considered the most vulnerable period, as a convincing fake video released then might not be debunked in time.
Potential sources of interference
Several actors could potentially interfere with the integrity of the upcoming Cypriot elections. Russia is a well-documented source of pro-Russian propaganda in Cyprus, particularly as the country has shifted toward the West since 2023. Analysts have noted similarities between recent Cypriot incidents and previous Russian disinformation campaigns in other member states.
Turkey and structures in the occupied north represent another possible source of influence, though there has been no publicly documented attribution of responsibility for specific actions. Additionally, Iran has been linked by some officials to a wave of cyberattacks targeting European shops and infrastructure. These threats demonstrate that the risk of interference comes from a variety of state and non-state actors.
Beyond state actors, private intelligence companies and domestic operators also pose a risk. Investigators in previous incidents identified private firms involved in disinformation, even when the ultimate payer remained anonymous. Anonymous paid advertising networks can also be used to amplify divisive issues like migration or corruption without being easily traced to a single foreign source.
Proposed solutions and next steps
Experts suggest that Cyprus does not need to invent new solutions but rather adopt models that have already worked elsewhere. One proposal is to create a small, independent team accountable to parliament to coordinate the police, intelligence services, and media authorities. This would mirror the Swedish model or the Electoral Commission in Ireland to provide a unified response to hybrid threats.
Other recommended measures include establishing a specific protocol for the week between election rounds to notify candidates of deepfakes within hours. Cyprus also needs to pass legislation to implement the EU AI Act, which would include penalties for failing to label AI-generated content. Finally, political advertising should be monitored in real-time during the campaign rather than being audited years later.
Timing is essential for these changes to be effective before the 2028 vote. If the political decisions on these mechanisms are not made by March 2027, there will not be enough time to test them. The goal is to ensure that the government can provide immediate clarity to voters, rather than just analyzing the damage after an attack has occurred.
What to watch
- Whether Cyprus designates a national authority for the EU AI Act by the next deadline.
- The formation of a permanent, independent body to coordinate election security.
- Specific commitments from 2028 presidential candidates regarding the use of deepfakes.
Try it
Check the official Cyprus Radio Television Authority website to see what powers they currently have over online platform oversight.
Sources and links
- ReportingPhilenews: the article we sourced
Read the original at Philenews
First reported by Philenews. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.