The growing ease with which convincing fabricated images, audio, and video can be produced is quietly undermining a long-held assumption: that a recording captures something real. As synthetic media becomes more realistic and more accessible, the evidentiary weight that photographs, recordings, and footage have carried is eroding, with consequences that reach into journalism, the courts, politics, and the basic question of what people can believe.

For most of the modern era, a photograph or recording served as a form of proof. Although manipulation was always possible, it required skill, time, and resources, and the difficulty of producing a convincing fake meant that, in most circumstances, a recording could be taken as evidence that an event occurred. This assumption underpinned much of how societies establish facts, from news reporting to legal proceedings to the everyday trust people place in what they see and hear.

The tools that generate synthetic media are dismantling that assumption. Systems that can produce realistic images of people who do not exist, place real people in situations that never occurred, or fabricate audio of someone saying things they never said have grown more capable and more widely available. What once required specialized expertise can increasingly be accomplished with accessible tools, lowering the barrier to creating fabrications convincing enough to deceive.

The consequences extend in two directions. The first is the obvious danger that fabricated media will be used to deceive, to spread false information, damage reputations, manipulate opinion, or commit fraud. A convincing fake, spread quickly, can cause harm before it is debunked, and the debunking may never reach all who saw the original. The capacity to fabricate convincing evidence places a powerful tool in the hands of those who would use it to mislead.

The second consequence is subtler and perhaps more corrosive. As people become aware that any recording might be fabricated, the credibility of all recordings suffers, including genuine ones. A real recording of an actual event can be dismissed as a possible fake, giving those captured doing something they wish to deny a ready means of casting doubt. This erosion of trust in evidence, sometimes called the liar’s dividend, threatens to undermine the shared factual basis on which public discourse and accountability depend, as the mere possibility of fabrication provides cover for denial.

Efforts to respond have focused on detection and authentication. Techniques to identify synthetic media seek to spot the subtle artifacts that fabrication may leave, though this is a moving target, as the methods of generation improve to evade detection. Approaches to authentication work from the other direction, attempting to establish the provenance of genuine media through technical means that record where and how a recording was made, so that authentic material can be verified rather than relying on the detection of fakes. Neither approach offers a complete solution, and both face the challenge of keeping pace with rapidly improving generation.

The deeper challenge is social as much as technical. Establishing what is true has always depended on more than recordings, relying on corroboration, context, and trusted institutions, and the erosion of recordings as proof may require renewed emphasis on these other foundations of credibility. How societies adapt to a world in which seeing is no longer believing, and how they preserve the capacity to establish shared facts, will be among the consequential challenges posed by the technology, touching not only the integrity of information but the trust on which collective life depends.