Some decisions can create ripple effects. A great example of this is Meta’s latest initiative to curb the spread of misinformation through artificial intelligence. Although the company chose to beat the rise of artificially created content with its own AI software, the project resulted in catastrophic disappointment. Meta’s ‘made with AI’ label indicates such a disaster. The instrument mislabeled ‘real’ photographs as artificial, leading to a monumental backlash from photographers. And how did they correct their course? Well, through its ‘AI info’ tag.
The lead image is a screenshot of Meta’s ‘AI info’ tag, which was taken from the company’s blog for representation purposes.
It is not a joke. Meta seriously devised an ‘AI info’ tag to provide “transparency” and “additional context.” Per their updated blog, “Like others across the industry, we’ve found that our labels based on these indicators weren’t always aligned with people’s expectations and didn’t always provide enough context.” The new change “is now the better way to address manipulated media and avoid the risk of unnecessarily restricting freedom of speech.”
But how is Meta’s ‘AI info’ tag different from the last? The label will reveal whether an image was edited via software or if someone used a prompt. Additionally, Meta will utilize its “network of nearly 100 independent fact-checkers” to review false and misleading AI-generated content.
Since this circus commenced, many photographers have been searching for alternatives, disillusioned by the company’s lack of foresight. However, some are genuinely worried about Meta’s intentions. “Of all the companies I could pick to do the right thing for the right reasons, I’m pretty certain that Meta wouldn’t be on the list,” says street photographer Paul Harrison in an interview with The Phoblographer. “As a street photographer, an essential element of the art form is that photographs should be candid. Whatever happened happened. AI poses a threat to that. That said, it’s not as if it’s a new threat, with Photoshop being around for over 30 years.”
In the same vein, Meryl Meisler, who has been documenting the queer spaces of New York since the 1970s, emphasizes the challenges of Meta using a photographer’s work to train its AI. “As a photography-based artist, Instagram has become a platform to share my work with not just friends and followers. It has introduced my work to curators, journalists, and collectors. Now, I feel it is ‘risky’ and post less. I need to consult with my intellectual property attorney to understand the full impact and implications,” she tells us.
Although an initiative such as Meta’s ‘AI info’ tag will make people more vigilant and lead them to question the integrity of photographs that appear on their glowing screens, it will also sow the seeds of mistrust. “AI is only going to be increasingly integrated into photo editing software,” Paul reminds us. “What if a photojournalist used AI to remove dust from a film scan, and then they find their reputation trashed by Meta, potentially labeling their work as fake?” Paul’s question is pertinent during this moment as the lines between AI and reality have been blurring at an accelerated speed.
While photographers can discern the distinction between ‘editing’ and ‘manipulation,’ the general audience will never comprehend it due to the lack of knowledge about the medium. To them, ‘editing’ equates to concealing the truth. “Retouching has been a regular part of the photography process, used for both commercial publication and ‘fine art'”‘ purposes for decades, possibly since the invention of photography. Even with analog, traditional darkroom gelatin silver prints, we burn, dodge, and touch up dust spots,” states Meryl. “Photoshop and other imaging software replicate analog retouching tools,” she reminds us.
While Meta has made Instagram a monopoly in the photography business, risking the livelihood of many artists. “Facebook, Instagram, and other platforms were a free way to communicate and get exposure,” explains Meryl. “Now, we have become addicted users at a price of greater intellectual property infringement, or we must go ‘cold turkey’ and make our accounts private or delete them altogether.” There is also the growing apprehension that Meta is making the platform video-centric (to surpass TikTok), which means algorithms diminish the reach of photographers’ posts to their audiences. Unless, of course, they pay Meta to boost the images.
This brings us to the question: Will ‘straight’ and documentary photography prevail as opposed to experimental artworks? While we don’t know the correct answer, it certainly is making artists think about their approach. “It makes me want to become more of a purist, using my eye like a simple arrow instead of an assault rifle with rounds of endless ammunition,” says Meryl.
But we all eagerly await the question: Can Meta right its wrong in an ethical manner? Both Meryl and Paul think they can.
“A danger for Meta is that if they are trying to protect their service from bad actors, surely it’s the easiest thing in the world to trick their algorithm,” says Paul. “Anyone who already has the motivation to deceive people with AI will have the motivation to take steps to get their photo past Meta’s algorithm. I think I read that removing Meta’s new labels from your post is possible. Doesn’t that make the whole idea redundant?” he adds. Paul advises Meta “to stay out of this until they have a much better system in place.”
Similarly, Meryl states that Meta, one of the biggest conglomerates, should intelligently differentiate fabricated images from real ones. “If Meta easily distinguishes images that employ AI for retouching purposes, it can and should identify those created for fake news, propaganda, terrorist, election-swaying, and other nefarious purposes,” she states. “For example, how about Meta instituting ‘Fake News,’ ‘Propaganda Purposes,’ ‘Terrorist Tactic,’ ‘Supremacy Warning,’ and ‘Voter Suppression Info’ tags?”
This incident reminds us that for every question, there is a solution. It, however, begs the question: who will the answer benefit? Those in genuine need or those famished for more power?