Handling crises when the algorithm is to blame

It’s Saturday morning and I’ve not yet had my first coffee of the day. My telephone vibrates and a message from a colleague ensures there’s going to be nothing routine about that day. On the screen, the short text reads: “This is circulating. Have you seen it?” and is enough to take my breath away. A video follows. The image is clear, the voice familiar and the scenario seems credible. The person speaks with the calmness of those who know what they’re doing. And, for a few seconds, the mind accepts that as true — not because we’re naïve but rather because our brains have been trained down the years to trust in what we see and hear.

The video does not take long to break out of a closed circuit and enter the public domain. When it reaches journalists, it already comes with interpretations. When it reaches members of staff, there is already indignation. When it reaches partners and clients, difficult questions arise. And, thereafter, the clock changes speed: time shrinks, pressures rise and the scope for error disappears. Whoever is in a leadership position recognises the patterns: this is not just an issue in the media; this is a crisis situation.

In recent years, there has been much talk about artificial intelligence (AI) and with due reason given how this has simplified the production of convincing content. AI has cut the price of producing plausible texts, images and even videos but it does not shoulder all the blame: disinformation did not arrive with AI — and the most serious impact is not only on technology. This extends to the ways that disinformation, no matter where it comes from, alters the rules for crisis management: accelerating media cycles, fragmenting audiences, rending the context more adversarial and forcing decisions when still facing uncertainties.

The geopolitical turbulence of recent years is the factor responsible for certain changes in the paradigm. In the war between Russia and the Ukraine, for example, the information dimension has been handled as an integral facet to the conflict: narratives to justify actions, to discredit the adversary, to foster confusion and divide external supporters, this is all a deliberate act, persistent and objective oriented.  In the Gaza Strip, the speed and scale of flows of information — and disinformation — illustrate another facet of the present: conflicts, to a significant extent, take place online, with out of context images, contradictory statements and amplification campaigns on platforms where corrections rarely attain the same reach as the initial impact. And in the United States, disinformation has become a structural factor to internal tensions with direct impacts on the trust in democratic processes and in social stability. Even when the facts are verified, doubt persists as a political and psychological tool.

What does this actually change for those managing organisations? This, above all, shifts the nature of “incidents”. A disinformation crisis rarely emerges as an isolated piece. This begins with content (whether a video, audio or a screenshot) but which rapidly turns into a dispute over intentions, character and credibility. And this almost always emerges accompanied by a second layer: clips, “close sources”, biased interpretations, implicit accusations and questions designed to force a rushed answer. In these circumstances, crisis management is no longer an exercise in “finding the right phrase” and turns into an exercise in commanding. The first step is not to react — it is to organise. Who can confirm the facts? By what deadline? According to which sources? Who is talking to who? Who is going to answer to the journalists, to the regulators? Who is going to reassure critical clients and clarify the situation to members of staff? If this is not minimally prepared, what happens is predictable: unaligned messaging, declarations at the wrong time, internal contradictions and, in extreme cases, the actual organisation amplifying the problem.

There is an important difference between “deny” and “contain”. Denying involves saying that something is false. Containing is preventing an organisation be pulled along by the logic of manipulative content. Containing very often implies an initial short and sober response — sufficiently present so as not to seem absent and sufficiently prudent so as not to crystalise an error. In a crisis, the first communication serves above all to gain time with rigor: recognising that the theme exists, indicate that this is under verification, explain the process and commit to updating within a brief timeframe. This discipline is lower in profile but tends to be more effective than an emotional reaction. However, at the same time, there is a frequently underestimated front: inside the house. An organisation can lose control over a disinformation crisis not due to failing externally but due to leaving an empty space internally. If collaborators lack any context and orientations, seeking out wherever they can find this quickest — then this opens the doors to interpretation, rumours and involuntary sharing. A simple internal clarification, direct and human, with practical instructions (“do not share”, “submit requests by channel X”, “updates available here”) is, very often, the difference between containing a crisis and a crisis with its own combustion fuel.

Then comes the most difficult task: managing stakeholders in an environment in which nobody has much time and everybody has a great deal of anxiety. Regulators and authorities do not like surprises, investors do not like prolonged uncertainties; strategic clients do not like silence; partners do not like contradictions. Effective responses are rarely only public. They demand direct contacts, consistent and based on the same set of facts — even if this set is, at the beginning, small. And this demands clear language: avoiding repetition of the lie in trying to negate it and avoid certainties that later turn against the organisation. This is about avoiding promising what is not controlled.

There is also another point that the prevailing geopolitics has rendered more sensitive: the coordination and the origins of narratives. Not all disinformation is spontaneous and not all is domestic in origin. In periods of tension, the international events may be exploited by local actors to deepen divisions and distrust. Thus, for such reason, there is always the need for a constant methodology that spans competent monitoring, reading the patterns, the capacity of distinguishing organic criticism from coordinated amplification and prior relationships with relevant interlocutors — including journalists — prior to the crisis breaking out.

Finally, the message for those who manage and decide is practical and not at all theoretical: in the current context, disinformation is a crisis variable like any other — with the particularity of moving faster than the facts. Investing in the capacity for crisis management applied to disinformation is not “communication” in the light sense of the term; it is the protection of continuity, reputation and trust. Because when deceptive content reaches the top of the feed, what decides the outcome is neither indignation nor eloquence. It is the readiness: organisation, sobriety, coordination and the capacity to maintain credibility while the rest of the public space goes about losing its head.