Lancaster is a well-known teacher of library and information science. In this, his latest book, he and Warner report on a research project sponsored by the American Library Association. The project’s purpose was to examine the literature of artificial intelligence (AI), expert systems, and related technologies and to “advise the information service community on what can be applied today and what one might reasonably expect to be applicable to library and information services in the near future” (p. 1).

The project’s first tasks were to acquire the relevant literature through database searches; to examine the many hundreds of items that were retrieved; and to follow up by requesting additional information about items, products, and services that seemed of particular interest. The next set of tasks, after digesting the literature, was to identify library functions, such as descriptive cataloguing, subject indexing, acquisitions, and reference services, for which expert systems have been developed and applied. The general conclusion reached is that “Systems that have progressed to an operational state--i.e., are functioning on a daily basis and providing real service to library staff or users--are almost non-existent” (p. 40). The study next reviewed applications of intelligent technologies in fields other than libraries, such as education, health care, finance, and machine translation, where some forms of AI or expert systems are already being used. It was, therefore, assumed that these technologies could also be applied to the information service field, if not today, then in a digital library environment of the near future.

The last section of the book is titled “Implications for Managers of Libraries and Other Information Services.” The first implication is that these managers must keep informed about relevant new developments and products. Second, managers of individual libraries are cautioned against trying to design and implement their own AI/expert systems. Rather, the authors suggest that such a task could best be undertaken by a consortium of libraries or by a professional association. Finally, they point out that to function as a true expert, the system would need more than a large database of recorded knowledge and inference rules, because expertise requires evaluation and interpretation. Currently, even the most technically advanced AI and expert systems have limited capabilities.

This research study achieves its purposes of identifying and reviewing AI and expert systems projects in libraries and related fields, and advising the information science community about currently available technology and what can be expected in the near future. It would be reasonable to conclude that true expert systems are not yet available, but that useful systems are available and more are being developed.