Over the years, Five Books has become renowned among bibliophiles around the world for its curated book recommendations. Recently, the team at Five Books decided to interview ChatGPT, an advanced AI language model, and asked it to recommend the best books on Artificial Intelligence. To everyone’s surprise, ChatGPT not only provided the recommendations but also offered insightful justifications for its choices. This interaction raised intriguing questions about the role of AI in book curation.
To explore this further, the author of this article interviewed other Large Language Models (LLMs) such as Google’s Bard and Microsoft’s Copilot for their recommendations on books about India. The results were impressive, with several overlapping recommendations across the platforms. Books like Ram Chandra Guha’s India after Gandhi, P Sainath’s Everybody Loves a Good Drought, and Amartya Sen’s The Argumentative Indian appeared on all the lists. However, there was one unexpected addition: Suketu Mehta’s This Land Is Our Land: An Immigrant’s Manifesto, recommended by ChatGPT.
When asked about the inclusion of Mehta’s book, ChatGPT offered a compelling justification, highlighting its global narrative coupled with Mehta’s Indian heritage. The AI model emphasized that the book provides profound insights into the immigrant experience, offering a perspective that resonates with contemporary India. This reasoning resonated strongly, leading the author to consider the possibility of AI tools curating book, film, and music lists without human reviewers.
However, this idea faced resistance from various quarters. One colleague questioned the value-add of AI-generated recommendations, highlighting the absence of diverse perspectives and the presence of algorithmic biases. The author also acknowledged that a full-time Indian Tamil language writer, with greater exposure to Indian languages, might offer a completely different list. Furthermore, researchers working on LLMs at IISER pointed out that while these models excel in coding, they often struggle with writing research reports and papers, often fabricating citations and information.
The ethical concerns surrounding LLMs were also addressed. As they currently operate, these models do not generate original thoughts or opinions but instead reflect the consensus or dominant narrative found in their pre-fed information. This poses the risk of sidelining niche, unconventional, or emerging perspectives that human experts would appreciate and include in their recommendations. When asked for philosophical book recommendations, AI models are likely to suggest well-known Western authors like Plato or Nietzsche, while overlooking insightful but lesser-known contemporary thinkers.
In conclusion, the role of AI in book curation sparks both excitement and concern. While AI models like ChatGPT offer impressive recommendations, their limitations become apparent when considering the absence of diverse perspectives and potential biases. Human expertise and the ability to appreciate niche perspectives remain invaluable in the curation process. As a result, platforms like Five Books, along with experts like Badri Seshadri, continue to play a vital role in providing well-rounded and insightful book recommendations.
