Challenges and Reflections on the Practical Applicationof Large Language Model in Digital Humanities Work atLibraries

Abstract

Abstract:

The large language model has posed significant challenges to traditional knowledgecomprehension, processing, and services. The article employs literature review, expert interviews, andgrounded theory to conduct content analysis of expert interview data from four aspects: the value ofChatGPT’s digital humanities tools, the Chinese language corpus within large language models, theconstruction of Chinese language corpora in libraries, and the relevant ethical challenge. These analysesdelve into the challenges that large language model introduces to library digital humanities work. Basedon literature, practical scenarios, and expert insights, the article elaborates on the in-depth considerationsfor the practice of digital humanities work in libraries, including data ownership and copyright protection,language and cultural differences, biases, technological application boundaries, as well as standards andregulations.

Zhang Hongling, Shen Lili, Han Chunlei, Fu Yaming ( Shanghai Library, School of Cultural Heritage and Information Management, Shanghai University). Challenges and Reflections on the Practical Applicationof Large Language Model in Digital Humanities Work atLibraries[J]. Libraly Journal, 2023, 42(391): 31-39.

References

[ 1 ] 王静静，洪贇，叶鹰 . GPT 型技术应用重塑数字人文探讨[J]. 情报理论与实践，2023，46（6）：43–46.

[ 2 ] Bubeck S， Chandrasekaran V， Eldan R， et al.Sparks of artificial general intelligence： earlyexperiments with GPT-4[EB/OL]. [2023-05-25].https://arxiv.org/abs/2303.12712.

[ 3 ] 王金林.“诸神之战”：新工具抑或新存在[J]. 探索与争鸣，2023（5）：33–35.

[ 4 ] M Shanahan. Talking about large languagemodels[EB/OL]. [2023-05-25]. https://arxiv.org/abs/2212.03551.

[ 5 ] Raffel C， Shazeer N， Roberts N， et al. Exploringthe limits of transfer learning with a unified textto-text transformer[EB/OL]. [2023-05-29]. https://arxiv.org/abs/1910.10683.

[ 6 ] Liying Cheng， Xingxuan Li， Lidong Bing. IsGPT-4 a Good Data Analyst?[EB/OL]. [2023-05-29].https://arxiv.org/abs/2305.15038.

[ 7 ] OpenAI.GPT-4 Technical Report[EB/OL].[2023-05-25]. https://arxiv.org/abs/2303.08774.

[ 8 ] Rombach R， Blattmann A， Lorenz D， et al.High resolution image synthesis with latentdiffusion models[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and PatternRecognition（ CVPR） . New Orleans， LA， USA：IEEE， 2022：10674–10685.

[ 9 ] Lexis+AI ? Transform Your Legal Work [EB/OL]. [2023-08-13]. https://www.lexisnexis.com/enus/products/lexis-plus-ai.page?int-camp=legal-aibutton1-H01#contact.

[10] Scopus AI： Change the way you view knowledge[EB/OL]. [2023-08-13]. https://beta.elsevier.com/products/scopus/scopus-ai.

[11] Search Powered by Generative ArtificialIntelligence[EB/OL]. [2023-08-13] https://clarivate.com/news/clarivate-launches-enhancedse a r ch- powe r ed-by-gene r a t ive - a r t i f i c i a l -intelligence.

[12] AI学术研究助手（内测版）[EB/OL]. [2023-08-13]https://aiplus.cnki.net.

[13] 蔡子凡，蔚海燕 . 人工智能生成内容（AIGC）的演进历程及其图书馆智慧服务应用场景[J]. 图书馆杂志，2023，42（4）：34–43；135–136.

[14] 赵亮 . ChatGPT 及其对图书馆行业的影响[J]. 竞争情报， 2023， 19（3）： 43–44.

[15] 张智雄，曾建勋，夏翠娟，等 . 回应AIGC 的信息资源管理学人思考[J]. 农业图书情报学报，2023，35（1）：4–28.

[16] 陈云松 . ChatGPT 对人文社科研究的价值与挑战[J]. 探索与争鸣，2023（5）：13–15.

[17] 生成式人工智能服务管理暂行办法[EB/OL].[2023-08-13]. https://www.gov.cn/zhengce/zhengceku/202307/content_6891752.htm.

[18] Texts adopted：Artificial Intelligence Act[EB/OL].[2023-08-13]. https://www.europarl.europa.eu/doceo/document/TA-9-2023-0236_EN.html.

[19] IFLA Statement on Libraries and ArtificialIntelligence [EB/OL]. [2023-08-13]. https://repository.ifla.org/handle/123456789/1646.

[20] OpenAI， Terms of Use [EB/OL]. [2023-06-11].https://openai.com/policies/terms-of-use.

[21] Philip Wang. 以全球主流AIGC 产品用户协议为例梳理AIGC 生成内容的权利归属与使用限制[EB/OL]. [2023-06-06] https://www.lexology.com/library/detail.aspx?g=e4de09be-8be9-444b-b468-b1d8db37b906.

[22] Midjourney， Terms of Service [EB/OL]. [2023-06-11].https://docs.midjourney.com/docs/terms-of-service.

[23] 蓝江 . 生成式人工智能与人文社会科学的历史使命——从ChatGPT 智能革命谈起[J]. 思想理论教育，2023（ 4）：12–18.

[24] Ouyang Long， Jeff Wu， Xu Jiang， et al. TrainingLanguage Models to Follow Instructions withHuman Feedback [EB/OL]. [2023-05-26]. https://arxiv.org/abs/2203.02155.

[25] OpenAI. Democratic Inputs to AI [EB/OL]. [2023-05-26]. https://openai.com/blog/democratic-inputsto-ai.

[26] 托马斯·库恩 . 科学革命的结构[M]. 金吾伦