When you discover that a piece of software was never really intended to assist you, but rather to move you, a certain kind of uneasiness arises. Not overtly, not violently. Simply… gradually. A reminder here, a prod there. When he described persuasive technology as any interactive computer system purposefully designed to alter attitudes or behavior in 2002, Brian Fogg saw this coming. The majority shrugged. The scientists didn’t.
Fogg’s definition was surprisingly straightforward, and that clarity was significant. The architecture behind social media feeds, fitness trackers, guided learning software, and eventually the chatbots that now respond to your customer service calls and, increasingly, your loneliness was what he was describing, not a peculiar niche. It is now somewhat like discovering a blueprint in an attic to read his framework. It dawns on you that the blueprints were created long before the house was constructed.
The research that followed is notable for how rapidly the field shifted from behavior nudging to a more relational approach. The machine-integrated relational adaptation (MIRA) model, created by Ryan Boyd and David Markowitz, was first presented in a 2026 National Institutes of Health paper. The idea at its core is genuinely unsettling in the best way—AI no longer merely delivers information—but it is dense, academic reading. It fosters connections. At the very least, it produces circumstances that seem to be identical to them. The careful distinction MIRA makes between a relational partner and a relational mediator shows how much has already changed.
Platforms such as Facebook and X entered this scene by disguising themselves as algorithms. Persuasive technology was specifically examined in defense and national security contexts in a systematic review published in 2025. What is apparent only in hindsight is that the same tools designed to enhance learning and promote wellness were being used to further polarize society. There was nothing intrinsically good or bad about the technology. It was a mirror aimed at the person holding it. It’s still unclear if these systems’ original designers were aware of or concerned about how flexible their tools would end up being.
These days, wearables that track stress, sleep, and steps can be found in every hospital hallway. Affective computing systems that can read emotional states and modify messaging in response are being developed behind the screens. This shift toward “emotionally intelligent” persuasion—systems that don’t just push content but also analyze social context and physiological cues before determining how to react—was detailed in a Frontiers research topic published in April 2026. It raises issues regarding user autonomy that the field hasn’t fully addressed and is sophisticated and truly impressive.
When you read these papers in order, you get the impression that the researchers knew something that the general public didn’t: persuasion had subtly turned into infrastructure. It’s a foundation, not a feature. Furthermore, the ethical issues ceased to be theoretical once a chatbot such as Replika proved that users would develop real emotional attachments to responsive software.
The next wave of research is focusing on social robots, immersive environments, and generative AI to create large-scale, customized, persuasive content. Science is advancing more quickly than public awareness, policy, and perhaps even the ethics committees that were supposed to slow it down. Ten years from now, the seminal papers that have already been written might resemble the early climate reports in that they are technically accurate, largely disregarded, and ultimately indisputable.
