Integrated AI Assistants are built into trusted academic databases, publisher platforms, and library resources to help researchers find, summarise, and interact with scholarly content more efficiently. They can improve access to licensed and paywalled materials provided by the library.  As AI continues to evolve, we will likely see these assistants become standard features across library search tools.  

Key Shared Features

• Natural Language Searching: Most AI assistants use natural language searching rather than keywords, offering an alternative to find relevant material.
• Accuracy: AI generated responses are grounded in the database, reducing the chance of bias and hallucinations.
• Data Protection: As library databases adhere to data privacy policies and regulations, these offer a safer search space than other types of AI research tools. 
• Environmental Impact: AI assistants are limited to searching within the content provided by the database itself, rather than trawling the web, so they are likely to be relatively sustainable. They also often rely on existing Large Language Models (LLMs), which helps them to be more efficient.

Limitations

• Search Scope: Limited to searching content or, even a specific item, within the database.