Bias in Data and Interpretation: AI systems rely on datasets that may contain historical biases, omissions, or skewed perspectives. If the training data is incomplete or reflects dominant narratives, AI might perpetuate inaccuracies or marginalize underrepresented voices. For example, colonial-era records often underreport indigenous perspectives, and AI could inadvertently amplify this issue by prioritizing such records in its analysis (Noble, 2018).

Lack of Contextual Nuance: History requires interpreting complex human motivations, cultural contexts, and ambiguous events. AI may struggle to capture these subtleties, potentially oversimplifying or misrepresenting historical events. It might prioritize patterns in data over qualitative insights, leading to reductive conclusions that fail to account for the multifaceted nature of historical narratives (Guldi & Armitage, 2014).

Accountability and Transparency: Who is responsible for AI-generated historical narratives? If an AI produces a flawed or misleading interpretation, it’s unclear whether the developers, data curators, or users are accountable. Additionally, opaque algorithms can make it hard to understand how conclusions were reached, undermining trust in AI-driven historical analysis (Mittelstadt et al., 2016).

Manipulation and Misuse: AI could be used to rewrite or distort history for ideological or political purposes, such as generating revisionist narratives or deepfake historical content. This risks spreading misinformation, especially if outputs are presented as authoritative, potentially influencing public perception of historical events (Paris & Donovan, 2019).

Dehumanization of Historical Inquiry: History is a human endeavor, shaped by empathy, moral reflection, and debate. Overreliance on AI might reduce historians’ roles, sidelining the human judgment needed to grapple with ethical questions or conflicting sources. This could diminish the interpretive depth that human historians bring to the field (Australian Historical Association, 2021).

Access and Representation: AI tools may be developed by specific groups or corporations, potentially limiting access to diverse scholars or communities. This could concentrate historical narrative-building in the hands of a few, excluding marginalized groups from shaping their own histories and reinforcing existing power imbalances (D’Ignazio & Klein, 2020).