AI does history

I’m revisiting older blog posts. I’m up to the one titled “Loose ends,” which reflects on the nature of origin stories in the age of the Internet.

The post from 2013, mentioned horsemeat detected in hamburgers, Derrida on the desire for beginnings, and Freud on red threads in naval ropes.

I have recently traced a different personal origin challenge: the story of my father’s wartime service. I recruited ChatGPT in this task, a consultation that took the form of a conversation.
In working through my father’s wartime story, we moved step by step from scraps of evidence — photographs with brief labels, a set of medals, and family anecdotes — towards a plausible account of his service.

In its analysis of a photograph of a group of men relaxing in a river, the software suggested that “the landscape and wartime deployments point strongly to the Brahmaputra valley in Assam, with the likeliest places being Guwahati, Tezpur, or Dibrugarh. Those were key river sites where British troops transited, trained, or rested during the Burma campaign.”

Another image shows men standing before a jeep in the desert. I also uploaded a photograph of my father’s medals. The AI was able to identify the location of the desert photograph: “The medals confirm the Africa Star and the Burma Star, and the phrase ‘the forgotten regiment’ recalls the ‘Forgotten Army’ in Burma. Slowly, a narrative takes shape: a young man from Mossley conscripted perhaps at eighteen, serving in North Africa, fighting through Assam and Nagaland, then ending up in occupied Berlin.”

Out of this process emerged a narrative linking North Africa, Burma, and post-war Berlin. I was aware that this was a generative reconstruction, a weaving of threads, a hypothetical sketch rather than an authoritative account.

The evidence was fragmentary: a few photographs labelled “India during the war,” a group of medals, memories of a scar from a motorbike accident, family stories about Berlin after 1945. Piecing them together was like following threads through a rope, tugging at details to see what connects and unravels.

The next steps would require verification: consulting official service records through the National Archives, checking regimental histories, and perhaps contacting the Manchester Regiment Museum. These would refine, correct, and lend stronger evidential support to the story that ChatGPT helped me to sketch.

Like the “loose ends” post I wrote in 2013, these threads hold together through the weave of a story. ChatGPT summarised: “the narrative is provisional, never fully tied off. Origins resist neat resolution: whether of hamburgers or family histories. What matters is the weaving — the act of tracing connections across fragments, texts, photographs, memories. That is what produces meaning, and perhaps what binds us to those who came before.” I would add in Freudian or Derridean mode the possibility that such investigations could also unbind us from what came before.