
BBC Question Time has used AI-generated versions of historical figures as part of a special episode about artificial intelligence, turning a current-affairs programme into a live case study in synthetic media disclosure.
The special was presented by Fiona Bruce and opened with an imagined AI panel of historical figures. A Radio Times preview named Mahatma Gandhi, Frida Kahlo, Che Guevara and Emmeline Pankhurst, while reports after transmission described the aired segment as featuring Winston Churchill, Gandhi, Kahlo and Pankhurst.
The BBC’s stated aim was to show audiences how hyper-real and persuasive AI-created images can be, and how easily they can blur the line between reality and fakery. That makes the production choice more interesting than a simple gimmick. The programme was not trying to pass the figures off as real, but to make the mechanism visible inside a mainstream broadcast.
For broadcasters, the useful issue is disclosure. Synthetic media does not only create risk when it is deceptive. It also creates editorial questions when it is clearly labelled: how prominent should the label be, how should presenters introduce it, what context does the audience need, and how far can a programme go before illustration starts to look like endorsement?
The segment also touches on likeness and representation. The figures used were dead, famous and historically loaded, which lowers some consent issues but raises others. Even when an AI recreation is framed as educational, it still asks viewers to accept a manufactured version of a person who cannot approve the performance, the framing or the words placed around them.
The live panel that followed included real political and technology voices, including Synthesia CEO Victor Riparbelli, putting the synthetic opening beside a more conventional debate about AI risk, public trust, jobs and regulation.
The sharper lesson for media teams is that AI disclosure is becoming part of production design. It is not enough to decide whether an image is synthetic at the end of the process. Producers need to decide why it is being used, how it will be labelled, what audience confusion is possible and whether the editorial purpose survives the obvious viewer reaction of “why did you make this slightly cursed thing?”