AI can generate compelling historical stories in seconds, but it cannot replace transparency. This article explores why provenance, accountability, and source attribution are becoming essential strengths of digital collections and why libraries and archives are uniquely positioned to provide trust in the AI era.
Recently, we stumbled across a social media post telling the story of a woman who allegedly pushed her four children 800 miles in a baby carriage from Missouri to Colorado in 1911 to secure a homestead claim. Like many compelling historical stories shared online, it had generated hundreds of comments and reactions.
Whether this particular story was AI-generated, embellished over time, or entirely authentic is almost beside the point. The reality is that we now live in a world where creating compelling historical stories, images, and narratives has become faster, easier, and more accessible than ever before.
At first glance, everything about the post appeared credible. It included exact dates, names, ages, distances, locations, newspaper-style headlines, and what looked like a period photograph.
Many of the surrounding details are also grounded in genuine historical context. The Homestead Act was in effect at the time, families did travel extraordinary distances to claim land, and women often faced immense hardship on the frontier.
Without a clear source, readers are left unable to distinguish between documented historical fact, interpretation, and details that cannot be independently verified.
Increasingly, that's becoming one of the defining trust challenges of the AI era.
The blurring lines between history and storytelling
For generations, historians worried about records being lost. Today, we face a different challenge: historical stories and records whose origins are not always clear.
AI can be a powerful tool for storytelling and bringing history to life, and many people are using it for entirely legitimate purposes, from education and genealogy to creative expression and sharing a passion for the past.
The risk arises when generated content becomes disconnected from its origins and begins circulating online without clear source information.
Which makes one question more important than ever: where did this information come from?
Provenance matters more than ever
Provenance is the documented history of where information originated.
Whether it's a newspaper article, birth certificate, census record, photograph, diary entry, map, oral history, or government document, provenance provides the context needed to understand, trust, and independently verify historical information.
This may include knowing which institution holds the original record, when and where it was created, how it was digitised and described, and the metadata and systems used to preserve and provide access to it.
In many ways, provenance is becoming one of the most valuable services libraries, archives, and digital heritage organisations can provide. AI can generate content, but it cannot replace the transparency that underpins trust.
The evolving role of digital collections
For many people, digital collections are synonymous with access and discovery. They allow researchers to easily search millions of records, uncover stories, and connect with the past.
In an AI-driven world, however, digital collections are important repositories of trust, providing something social media often cannot: accountability.
When researchers discover a historical record within a professionally managed digital collection, they can trace it back to its source, understand its context, and identify the institution responsible for preserving it.
That chain of evidence matters, and in the years ahead, it may become one of the defining characteristics that separates trusted historical resources from everything else online.
AI still has an important role to play
While libraries, archives, and cultural institutions continue to navigate important questions around copyright, licensing, and how collections may be used to train AI models, there is also a significant opportunity to thoughtfully embrace AI as a tool for improving access to history.
Related reading: AI Models and Copyright Concerns
When implemented responsibly, AI has the potential to help researchers uncover connections that may have previously gone unnoticed, surface relevant records more quickly, summarise large volumes of material, and create richer discovery experiences:
Its role, however, should be to enrich and build upon historical records, not separate them from their origins.
We believe that discovery should begin with the source
We've always believed that discovery should be grounded in honouring the source material.
The Veridian platform is designed to preserve the connection between search results and the institutions that hold and care for original records, allowing users to navigate from a discovery experience back to the source itself, whether that's a newspaper page, photograph, document, or other historical record.
As AI-generated content becomes increasingly sophisticated, digital collections have an opportunity to reinforce their value not simply as repositories of information, but as trusted stewards of historical knowledge.
Technology will continue to evolve. AI will likely become faster and more capable.
One thing is unlikely to change, however: the future of historical research will not depend solely on who can generate information the fastest. It will depend on who can demonstrate where that information came from, a role libraries, archives, and trusted digital collections are uniquely positioned to fulfil.