Chatbots as oracles: How AI is giving new life to the ancient practice of seeking answers from inscrutable sources

In recent years, people are increasingly seeking guidance from artificial intelligence (AI) systems. According to a 2026 Pew report, 10% of Americans rely on chatbots for emotional support or companionship. A large-scale study of 1.1 million ChatGPT conversations revealed that 49% of the messages were users seeking information or advice to enhance their knowledge or decision-making.

The questions posed to AI go beyond ordinary practical matters; for instance, ethical or moral concerns were a significant focus in 2021 when Ask Delphi, an AI-based website, received 3 million queries.

The name "Ask Delphi" evokes the ancient Delphic oracle, a priestess who entered a trance state to convey answers from Apollo, the Greek god of prophecy. This association has led some to perceive chatbots as inscrutable sources of superior knowledge. The Delphic oracle's power stemmed from her ability to provide guidance on questions that ordinary human judgment couldn't answer. Throughout history, societies have turned to oracles, prophets, and divination practices for answers when human judgment reached its limits.

The omnipresence of oracles across cultures and religious traditions highlights the universal human desire for transcendent sources of insight. For example, the Ram Shalaka oracle, based on the Hindu scripture Ramcharitmanas, and Christian bibliomancy, where a user opens the Bible at random, both demonstrate how randomness creates an enigmatic black box effect. This opacity reinforces the perception of oracles as possessing knowledge beyond human limits, regardless of the underlying mechanisms.

Alan Turing, a pioneer in computer science and artificial intelligence, introduced the concept of "oracle machines" in 1939. These mathematical models of computation could consult external sources for answers unavailable to the machine itself. In both mathematical and religious contexts, an oracle provides answers to problems beyond the scope of an existing system or set of axioms.

Modern chatbots like ChatGPT are considered opaque black boxes, even though the computations they perform are relatively simple. The automatic learning of weights from vast data during training further complicates the understanding of how these models arrive at specific answers.