In March 2026, a photograph began circulating on social media purportedly showing a McDonald’s ordering kiosk that had produced an absurd, offensive, or dangerous output — depending on which version of the image you saw.
The image was compelling. It was picked up by mainstream press. Fast Company covered it. The Chipotle Pepper wrote about it. Executive teams across the fast-food industry and adjacent sectors began asking whether their own AI-mediated customer interactions could produce similar liability.
By April 2026, Fast Company published a detailed follow-up establishing that the viral image had been substantially fabricated. A woman named Sally Evans, who had been quoted in the original circulation, gave a direct statement: “The viral post was Photoshopped.” The specific claims about kiosk behavior were not supported by the underlying evidence.
By that point, the story had already shaped executive perceptions of AI risk in ways that no correction would fully reverse.
The information layer
This is the first of three pieces on what I call the McDonald’s AI trilogy — three distinct layers of failure that senior leaders need to understand independently, because each requires a different response.
The first layer is the information layer: the environment in which senior leaders form their understanding of what AI is doing, what risks it poses, and what interventions are required.
The viral McDonald’s story is a specific example of a general pattern. In 2026, the information environment surrounding AI is saturated with:
Fabricated screenshots and images purporting to show AI failures.
Real AI failures being dramatically exaggerated in retelling.
Real AI successes being similarly exaggerated by vendors.
Consulting-firm case studies where the reported outcomes bear only loose resemblance to what actually happened at the client.
Podcast conversations where speculation is presented as reporting.
Conference keynotes where thought leadership is confused with substantive analysis.
The senior leader forming their understanding of AI in this information environment is not, in most cases, forming their understanding based on rigorous evidence. They are forming it based on a saturated mix of accurate reporting, dramatic misinformation, vendor marketing, and consultant self-promotion.
The specific failure mode
The failure mode created by this environment is not that senior leaders are being deceived by a single false story. The failure mode is that senior leaders are systematically miscalibrated about the distribution of AI risks and opportunities.
The dramatic failures they hear about are overrepresented. The mundane successes are underrepresented. The vendor claims are amplified. The independent research is not.
The result: executive decisions about AI investment, governance, and risk management are being made against a distorted picture of the actual situation.
Three practical questions
One: what specific AI-risk stories has your executive team discussed in the last six months? How many of them have been rigorously verified? How many have been treated as authoritative without verification?
Two: what is the source composition of your organization’s AI intelligence? Vendor-produced. Consulting-produced. Media-produced. Independent-research-produced. First-hand experimentation. If any single category dominates, the picture you are forming is not representative.
Three: what would falsify your current AI risk model? If nothing would falsify it, you are not operating on a model. You are operating on a collection of stories.
The closing thought
The information layer is the first layer of the McDonald’s AI trilogy because it shapes everything that follows. If you understand the operational layer or the architectural layer correctly but consume the information layer uncritically, your interventions will be miscalibrated.
The viral McDonald’s story of March-April 2026 was a specific instance of a general pattern that will recur throughout the next decade. Fabricated AI-failure stories will circulate. Some will be caught quickly. Some will not. Executive decisions will be shaped by both.
The discipline of information hygiene — verifying sources, checking claims, tracking corrections, weighing evidence — is the first layer of executive practice that AI transformation actually requires.
Parts two and three of this trilogy address the operational layer and the architectural layer respectively.
The world has changed. The leaders who notice will be the ones the next decade is built around.
