AI: Threat to Truth or Renaissance of Possibility?
A writer and academic, two lawyers, and a digital editor contemplate the future of photojournalism and truth in the context of Artificial Intelligence
by Maya Baylis
Can the public ever trust photographs again? The audience snapped to life as Friday’s first 4 panellists stepped up to tackle the unprecedented paradox of AI, both a threat and what Fred Ritchin described as a “renaissance of possibility”.
Anna Skurczynska set the tone in solicitor fashion by encouraging slow critical consideration when collecting information, “so that we are neither gullible nor numb to the point of not caring.” Skurczynska used the recent case of Charlie Kirk as an example of how the law is much more critical of evidence than the over-excited media.
Skurczynska suggests, even pre-AI, we never had access to the one truth: “I talk about facts, and not the truth. Because truth, in a way, is an aspirational idea.”
“There’s a real, existential threat to our work now, and to evidence-based journalism.” Sophie Nicholson of Agence France-Presse revealed how one of the main news agencies has adapted to use AI for in-house tools like translation, verification, or transcription, but never in what’s published. “With detection becoming more challenging, we hope to advance labelling with some kind of digital tattoo or watermark that can't be scraped.”
“We had to discontinue the case without the witness statements being heard”, says lawyer Jennifer Kanis, “this meant records of witness statements would be lost.” The Australian law firm’s project Exhibit AI unveiled AI’s potential.
Witness statements provided by asylum seekers held in notorious detention centres on Nauru and Manus Island were memorialised in collaboration with AI prompt-engineers. Yet, Kanis emphasised an important semantic distinction: “we weren’t trying to create photographs; we wanted to create images to show a true reality.”
“AI undermines the power of real evidence,” writer and academic Fred Ritchin says. In 1984, he was already concerned about losing the photograph as a credible witness to injustice. Now, the credibility is eroded by our inability to tell generative AI images apart from real photos.
Despite condemning many “obscene” uses of AI to rewrite history, Ritchin displays some of his own explorations with AI prompt-engineering – visualising the unphotographable: dreams, nightmares, martians, or the future. “I think we need to stop thinking of photography as the only visual depiction medium and AI for what it can do usefully […] It's not a binary, it's not good or bad, it's both.”
Aligning with tools suggested by AFP to “tattoo” AI-generated images, Ritchin pitched his Four Corner Project that conceptualises using text embedded in images. He urges everyone to see this as the start of a new non-fiction versus fiction imagery era.
Like photography forced painters to create cubism and impressionism, Ritchin hopes AI forces photographers to rethink the craft. “And I think the results will be coming from you guys.”
