Algorithmic Amplification Is Now a Critical Infrastructure Risk

Critical infrastructure incidents no longer unfold only across physical assets and cyber systems. They also unfold in an infospace shaped by digital platforms whose algorithms favour speed, emotional intensity, and visibility over verification. For operators, this means that rumour, distortion, and misleading narratives can become part of the incident environment before the first official statement is issued. That is no longer just a communications issue. It is a resilience issue.

The problem is not limited to deliberate disinformation campaigns by hostile actors. A significant part of the risk arises from the normal operation of large social media platforms. During an outage, transport failure, communications disruption, or industrial accident, there is almost always an early information vacuum. Official facts are incomplete, technical assessments are still underway, and restoration timelines remain uncertain. Into that gap flows the content platforms are most likely to amplify. In practice, that often means the most dramatic, emotionally charged, or speculative interpretation of events.

The consequences are operational, not merely reputational.

First, algorithmically amplified misinformation can distort public behaviour. False or exaggerated claims about the scale of an incident, the existence of secondary hazards, or the reliability of official updates can influence how people respond in real time. In an infrastructure emergency, resilience depends partly on public cooperation: passengers need clear guidance, consumers may need to reduce demand, and affected communities need to know which channels to trust. Confusion in the information environment weakens that cooperation.

Second, it can place additional strain on response systems. When misleading narratives spread quickly, emergency services, customer contact centres, and public authorities may face surges of enquiries, complaints, or panic-driven reactions that do not correspond to the technical reality of the event. This diverts attention and capacity at precisely the moment when disciplined prioritisation matters most.

Third, it can damage trust in ways that outlast the incident itself. Even where restoration is technically successful, the dominant online narrative may frame the event as evidence of incompetence, secrecy, or negligence. Once that framing becomes established, later incidents become harder to manage because the baseline level of trust has already been weakened.

For this reason, critical infrastructure operators should stop treating the infospace as a downstream issue to be handled only after the technical response is underway. In the platform age, the narrative environment develops in parallel with the operational incident, and often faster than formal communication processes can respond. The information dimension therefore needs to be built into resilience planning in advance.

Three practical steps would make a real difference.

The first is to treat narrative monitoring as part of situational awareness during significant incidents. Operators already monitor technical status, cyber indicators, and service impacts. They should also have a structured way to assess what claims are circulating online, which narratives are gaining traction, and whether those narratives are affecting behaviour. While this is established practice in advanced crisis communication teams, it remains rare in CIP incident management structures — and that gap needs to close.

The second is to invest in trusted communication channels and pre-prepared messaging frameworks before a crisis occurs. Credibility cannot be improvised in the middle of a major outage. Organisations that have already established visible, consistent, and recognisable public channels — whether through direct social media presence, relationships with local authorities, or pre-agreed communication protocols with emergency services — are in a much stronger position to reach the public quickly with accurate information when it matters. Equally important is preparing draft holding lines and narrative frameworks for the most foreseeable incident types: a grid outage, a transport disruption, a communications failure. When the information vacuum opens, the difference between responding in minutes and responding in hours is often the difference between shaping the narrative and chasing it.

The third is to update crisis training and exercises. Tabletop exercises should not simulate only the technical disruption. They should also simulate the platform dynamics surrounding it: rumour cascades, miscaptioned images, false attributions of cause, and competing unofficial narratives. That is now part of the real operating environment, and preparedness should reflect it.

Critical infrastructure protection has always required attention to the wider conditions that shape disruption. Today, one of those conditions is the infospace. Algorithmic amplification is not an abstract media issue or a secondary public relations concern. It is part of the environment in which incidents are interpreted, escalated, and managed. Resilience planning should reflect that reality.

Antonio Scala is a physicist and Research Director at CNR-ISC. His research focuses on complex networks, information dynamics, and critical infrastructure resilience.