'Using Generative AI in Historical Practice' argues that generative models are reshaping historical scholarship. Rejecting medium - and long -term speculation, it focuses on near-term practice: how historians can use AI now to augment their research through context-aware dialogue, semantic search, network visualization, multimodal source analysis, and code-assisted workflows. It details methods for context management, task design, and response structure, while warning against cognitive offloading and model bias. While it offers a variety of novel methodologies, the book insists on the indispensability of human agency and taste. Case studies range from Augustine of Hippo to early cinematography, demonstrating the possibilities and limits of generative AI. It concludes with a call to historians to engage with the technology critically and productively, reimagining AI-assisted scholarship without surrendering disciplinary standards and aims.
1 有用 (*☻-☻*) 2026-05-18 16:38:03 浙江
作者对ai对历史研究的影响还挺乐观的,“在可预见的未来”,专业历史学家的taste和agency还是不可替代。给ai任务,让ai自己写提示词、完成流程设计,分享了自己和ai交互的例子(甚至考虑了链接过期给了notion)。作者的主要态度就是,在保持警惕的时候接受ai融入历史研究已是大势所趋。
0 有用 吃蛋挞 2026-05-14 16:53:28 北京
一本几十页的小册子。废话太多,有些技巧可堪借鉴。
0 有用 Martinaise,R. 2026-05-29 14:34:39 浙江
"历史学家需要熟悉这些工具及其底层技术,才能有效地做到这一点。他们必须认识到生成式人工智能的陷阱和偏见,但更重要的是,他们必须理解其能力和方法论,以形成一个心智参照框架,这将有助于他们在研究中应用它。"——安全而合乎伦理地适用LLM;总而言之要在尽可能保留作为历史学家自身能动性的情况下主动拥抱ai趋势,因为已经势不可当。
0 有用 Res Novae 2026-05-29 20:17:36 英国
值得速读。可能是因为交稿原因,作者没有特别思考Agentic Use. 但整体上来说,对于semantic search和network的分析确实是当下比较好用的方向了。