91��ɫ

By, Victoria Rigas-DiDomenico

In today’s digital age, AI-generated images have become commonplace on social media. When scrolling various platforms, it is almost certain that you will come across at least one photograph, video, or written post that contains elements of generative AI technology. The rapid development of artificial intelligence over the past few years has fundamentally altered the media that we consume on a daily basis, impacting our ability to distinguish fiction from reality. 

In the field of history, AI has become a major topic of debate. Photographs depicting historical episodes, people, and spaces have become increasingly popular online, producing false narratives about the past. In this post, I will outline three main problems with AI-generated historical photographs: historical inaccuracy, ethical concerns, and disregard for primary sources. I will also suggest potential action points that should be taken into consideration moving forward. For better or for worse, the use of AI technology has become part of our lives, and seems to be here to stay. I hope that this reflection will provide a necessary commentary on what it means to be human in a digital world, advocating for critical analysis, digital literacy, and increased skepticism online.

Showcasing "Vintage Toronto?" or the Ultra-Modern?

Recently an image was shared on social media showing Greeks during a Good Friday procession on the Danforth in 1974. Although this photograph is AI-generated, it has gained over one thousand “likes” on Facebook, marking significant engagement from a wide audience. 

This photograph, while visually striking, presents many factual inaccuracies. On the surface, one can see a candlelit procession that appears to be taking place during broad daylight. In the Greek Orthodox tradition, the only procession that occurs during Easter takes place in the evening on what is known as Good Friday. This procession is distinct and involves the carrying of the Epitaphios (a physical manifestation of the tomb of Jesus Christ). In this photograph, the light sky and absence of the Epitaphios demonstrate clear markers of AI intervention.

We can also note many issues with regard to the storefronts that make up the background of this image. While this photo indicates that it was captured on Danforth Avenue, it shows businesses that did not exist sequentially or aesthetically as they are shown here. For instance, the exterior of the “Danforth Music Hall” is completely falsified. 

Perhaps the main giveaway for this photograph is the banner strung across the street that reads: “Greek Easter 1974 - Danforth.” The presence of this banner is an unrealistic inclusion to the cityscape, alluding to the fact that this photograph was AI generated. 

While this Facebook account declares that they are presenting a certain “Vintage Toronto,” one could argue that this “photograph” is more representative of an ultra-modern Toronto – the city as imagined by artificial intelligence. 

The Impact of AI on History

When it comes to photographs and the depiction of historical events, AI poses a major threat. The most prominent issue with AI photographs of historical events is the promotion of historical inaccuracy. AI-generated “photographs” do not capture real life moments and therefore misrepresent the past. These images are created with no historical evidence and therefore distort the complexity of what actually occurred. For viewers who might be unaware of the context of the image, they are likely to take what appears on the surface as fact, constructing an uninformed and inaccurate knowledge base. 

There are also major ethical concerns. AI images can be oversimplified and insensitive to the emotional and personal attachment that viewers might have to the moment pictured. For those with personal stakes in the events illustrated, it can be disorienting or offensive to have their lives misrepresented. 

One should also note that AI images are not “primary sources.” A primary source is a first hand account, often taking the form of written documents, photographs, or other art forms. Primary sources provide the key to unlocking knowledge about the past and assist historians in constructing historical narratives. AI generated images complicate primary sources due to their perceived “accuracy,” undermining scholarly authority, archives, and research. AI generated images are produced, not captured organically, and therefore do not accurately convey a particular moment in time. 

The main contention between AI and history is the potential to reinforce inaccurate or biased narratives about the past. As such, these images can dismantle years of work done by historians to construct plausible historical accounts. While it can be argued that all history presents some form of bias, or that all photographs can be edited after they are taken, the main issue with AI is how it deceives the viewer. AI generated images do not disclose that they are produced by AI, they do not provide a source base that is traceable for the viewer, and they do not outline how they may have manipulated various images to construct the final product. The writing of history should be transparent and traceable, a facet that is missing entirely from AI generated images – especially when posted on social media. 

What Can We Do?

While we cannot simply ignore or “get rid” of AI produced photographs, we must engage critically with what we see online. I encourage viewers to reflect upon the context, accuracy, and staging of these images, while also consulting other primary source documents of similar events. For instance, we can compare the AI generated photograph of “Greek Easter” with photographs from the Clara Thomas Archives at 91��ɫ. Take, for example, this image produced by the Toronto Telegram. 

This photograph is dated to 1968 and presents a credible example of a primary source. This photograph was produced by the press (Toronto Telegram) and is cataloged in a reputable archive. On the surface of this image, one can see an authentic example of a Good Friday procession, drawing a stark contrast to the image produced by “Vintage Toronto.” In this photograph, one can see a dark sky and presence of the physical structure of Epitaphios, demonstrating verifiable elements of this historical event. When primary source photographs are placed side-by-side with AI generated images, viewers are able to point towards potential discrepancies.

Final Thoughts

Now, more than ever, it is imperative that we consult primary sources and reflect upon whether an image reflects historical fact. Digital literacy is necessary to how we engage with social media and it is our responsibility to become informed about the potential risks of AI. We must be increasingly skeptical about what we see online, especially images claiming to represent a historical reality.

For more photographs related to Greeks in Canada – devoid from AI intervention – I invite you all to consult our physical and digital collections: /research/project/hhfgca/physical-holdings/

Further Reading:

Peer Reviewed:

Chang, Briankle G, and Brooke Belisle. “AI and Photography: An Interview with Brooke Belisle.” Photography & Culture 18, nos. 3–4 (2025): 305–11. https://doi.org/10.1080/17514517.2025.2488170.

Meyer, Roland. “Images from Images: Generative AI and the Reconfiguration of the ‘Photographic.’” Photography & Culture 18, nos. 3–4 (2025): 281–90. https://doi.org/10.1080/17514517.2025.2522517.

Ritchin, Fred. The Synthetic Eye : Photography Transformed in the Age of AI. Thames & Hudson, 2025.

Wasielewski, Amanda. “Unnatural Images: On AI-Generated Photographs.” Critical Inquiry (Chicago) 51, no. 1 (2024): 1–29. https://doi.org/10.1086/731729.

Articles:

Parkman, Chloe. “Historian Calls for AI Content of Occupation to Be Labelled.” BBC News, March 5, 2026. https://www.bbc.com/news/articles/crl44g70dklo.