Critical Thinking in the Age of AI

A screenshot appears in a chat and looks like breaking news. It might be a cropped headline with a “BREAKING” banner, or a repost with a new caption added. The language is confident and urgent. Within minutes, the image can do what used to take a full story or an official statement. People react to what they see, not what they can confirm.

Propaganda is messaging built to steer an audience toward a belief or response. It is designed to feel obvious, and it is repeated until it sticks.

World War II shows how influence can spread through repetition and presentation, often more than through detailed argument. One airpower-related example was the use of U.S. Office of War Information psychological warfare leaflets in the Pacific. These leaflets were designed to be dropped by air over Japan and Japanese-occupied territory, including from B-29 Superfortress bombers. The delivery method mattered because aircraft could place the message over areas that other media could not reliably reach. By the end of the war, tens of millions of leaflets had been dropped over Japan. Each leaflet carried a short message meant to shape how people understood the war and the choices available to them.

That is the goal of this type of messaging. When it shows up everywhere, it shapes what feels normal. Persuasion can support useful habits, including caution with sensitive information during a national effort. It can also narrow what people see by leaving out context or presenting one version of events as the only reasonable one. In military settings, influence can be aimed directly at morale and decision-making.

Operation Moolah during the Korean War is another airpower example. In 1953, U.S. Far East Command offered $100,000 for the first intact MiG-15 delivered to United Nations forces. The aircraft had major intelligence value because U.S. pilots had faced the MiG-15 in combat, and an intact aircraft could be studied and flight-tested. The offer also carried a message to enemy pilots: defection had a reward, and the other side was willing to receive them. No enemy pilot accepted the offer before the armistice, though a North Korean pilot later landed a MiG-15bis at Kimpo Air Base and learned about the reward after landing.

This history matters today because new tools make it easier to manufacture convincing “proof” for almost any storyline. Images and clips can be produced quickly, styled to look official, and shared as if they settle the question. Often, this approach is meant to shape first impressions fast and let the correction arrive after the moment has passed.

Today, people run into influence messaging in everyday places, including spaces that do not look political at first. This messaging shows up in social feeds, short videos, memes, screenshots, and group chats. It also appears through websites that mimic legitimate outlets and through edited clips that remove key context. The more shareable the format, the more it rewards immediacy over accuracy, which makes it easier for false narratives to be spread purposefully.

Some claims are spread through loose networks repeating the same message until it feels familiar. More deliberate campaigns are built to stir emotion and weaken trust. For a globally read publication, that matters because the same story can be packaged differently for different audiences and still feel believable in the moment.

The same pressure can now come from AI-generated or AI-altered content. A fake image can borrow the look of a newspaper, an official alert, or a broadcast clip, then move through social feeds before anyone checks the original. A correction may come quickly, yet the first version can still shape how people understand the event. That impression matters in a military-information environment, where a false image can create confusion around operations, aircraft, or partner forces.

One recent airpower-related case involved a fabricated image that appeared to show a Daily Telegraph front page praising the Pakistan Air Force. Reuters found no evidence that the newspaper published the headline, and other fact-checkers assessed the image as possibly or likely AI generated. The image borrowed the look of a familiar publication and attached that credibility to a claim about airpower. For readers moving quickly, the fake front page could feel like outside confirmation instead of another unverified post.

Another defense-related example shows how quickly an AI-generated image can travel. In May 2023, an image claiming to show an explosion near the Pentagon spread quickly online. The Department of Defense said there was no explosion, and Reuters reported that the image appeared to be AI generated. The false claim moved quickly because it looked like a breaking update involving a major defense location. A convincing image and an urgent caption pushed the story forward before verification caught up.

Across these examples, the pattern is consistent. When something looks credible and feels urgent, it creates pressure to react immediately, and that pressure can override good judgment.

The practical goal is to protect judgment in a fast and noisy information environment. Screenshots and cropped clips spread quickly because they hide where they came from, so treat an image as a lead, not as proof. Before sharing, pause and look for the original source and a fuller version of what you are seeing. Then, check whether a reliable outlet or official channel confirms the underlying claim. If all you can find are reposts of the same image with shifting captions, treat it as unverified.

Sharing is where good intentions can still spread false information. Reposting a fake image to criticize it, warn others, or ask if it is real still pushes it to new people. If you would not share it at work without checking, pause before sending it in a group chat. When you do need to correct misinformation, providing a reliable source or a verified correction instead of reposting the image itself is often helpful. Verify first, then decide whether it is worth amplifying at all. In a fast information environment, judgment is a readiness skill.