Why synthetic media is a new kind of cybersecurity threat

The threat synthetic media – digitally manufactured or manipulated content, from AI-generated video to voice clones – poses to democracy is no longer theoretical. During the ongoing 2026 US midterm cycle, political organizations have released realistic AI-generated videos depicting candidates saying things they never said, sometimes with disclosures that are easy to miss.

This shift is playing out globally: Brazil’s highest electoral court recently established binding standards covering synthetic depictions of political figures, while Reuters reports that AI content has become a standard campaign tool operating amid a patchwork of untested laws.

Traditional cybersecurity protects systems, networks and data. But synthetic media attacks the shared reality institutions use to make decisions. Against this backdrop, the question more and more governments and democratic institutions are asking is fundamentally operational: How can we establish what is real quickly enough to act?

That exact question surfaced repeatedly over the past few months during our briefings with election, security and public-sector leaders across the Americas, Europe, Asia-Pacific and Africa.

In fact, that urgent need for fast answers is already prompting real-world action. Ahead of the 2026 Scottish and Welsh elections, officials piloted procedures for detecting suspected deepfakes, and for rapidly notifying candidates, law enforcement and the public.

Yet generation is becoming faster, cheaper and more accessible, while verification remains fragmented and slow. Governments, platforms and news organizations each hold only part of the evidence. Closing that gap requires independent, multimodal technical analysis inside the communications and decision systems where consequential content first appears.

Without that rapid analysis, even a brief delay can have severe real-world consequences. Take, for example, a fabricated video of a military leader announcing troop movements, an audio recording purporting to reveal election interference, or a synthetic emergency message attributed to a national authority. Even if specialists ultimately prove them false, the fabrication may already have moved markets, triggered protests, altered military calculations, or undermined confidence in an election.

The most visible danger is that citizens will believe something fabricated. But the greater danger is that they may stop believing authentic information altogether.

Synthetic media gives bad actors two advantages: they can manufacture evidence and dismiss genuine evidence as fake. This “liar’s dividend” grows as the public becomes more aware of deepfakes. A 2026 audio study found that participants increasingly distrusted authentic speech even though their ability to identify synthetic speech had not materially improved.

This changes democratic accountability. A legitimate recording of misconduct may be rejected as synthetic. Authentic emergency instructions may be questioned precisely when rapid compliance is essential. Governments may hesitate to respond because officials cannot immediately determine whether the evidence before them is real.

This broader erosion of trust is accelerating. As the Reuters Institute’s 2026 Digital News Report warns, trust in news fell in 29 of the 48 markets they cover, driving global trust to its lowest recorded level since measurement began in 2015 – just 37%. In an information environment increasingly shaped by social platforms, AI chatbots and synthetic media, establishing whether digital content is authentic is becoming essential to restoring public confidence.

This decline reflects a broader systemic vulnerability. Recognizing the scale of the threat, the World Economic Forum’s Global Risks Report 2026 again ranked misinformation and disinformation among the world’s most consequential near-term risks.

Crucially, these technological and societal risks reinforce one another. Synthetic media becomes more effective in polarized societies, during armed conflict and wherever institutional trust is already weak. The technology does not have to create those divisions. It only has to exploit them.

No single intervention can solve this problem. Democracies need a layered trust infrastructure that combines authentication, independent detection, provenance, platform response, crisis protocols and governance. This framework relies on six interconnected pillars working in tandem to verify content and accelerate decision-making.

1. Authentication must validate official communications: Secure, publicly understood methods – including cryptographic signatures, verifiable credentials, and standards like C2PA – are essential to document origin and enable rapid confirmation during crises.

2. Detection must operate in real time: Independent, multimodal tools must continuously analyze anonymous or manipulated audio, video and images fast enough to inform high-stakes decisions before false narratives take hold. Recent audio research and the RADAR Challenge 2026 demonstrate both the urgency and technical complexity.

3. Provenance must document origin, where possible: Content credentials (such as SynthID or the C2PA standard) can provide valuable evidence of how media was generated or modified, but because bad actors easily strip or omit metadata, provenance cannot independently authenticate unsigned or anonymous uploads.

4. Platforms need direct escalation channels: Distribution networks must establish priority pathways to review and mitigate high-consequence synthetic content involving elections, public safety, military activity, or senior officials.

5. Crisis response protocols must define action: Clear inter-agency procedures must dictate who evaluates suspicious media, what evidence is required and how conclusions are communicated across agencies, allies, platforms and the public. These procedures should be exercised before elections and geopolitical crises, when ambiguity and delay can become risks themselves.

6. Governance must establish enforceable rules: Interoperable legal standards – such as the EU AI Act’s transparency requirements – should establish disclosure, evidence preservation and accountability requirements for synthetic content.

The central question is no longer whether synthetic media will be used against democratic institutions. It is whether these institutions are ready when it happens.

Platforms, media organizations, researchers and public agencies should collaborate to establish trusted communication channels before a crisis occurs, fostering shared responsibility for protecting democracy.

The objective is not to eliminate synthetic media or constrain legitimate uses of generative AI. It is to preserve society’s ability to distinguish evidence from fabrication when the consequences are greatest.

For much of the digital era, cybersecurity focused on protecting access to systems and information. The next era will require us to protect confidence in reality itself. Whether democratic institutions succeed will depend on whether the systems governing trust can move as quickly as the technologies now capable of attacking it.