Am I Being Gaslit by AI? A Critique of the Hype

Summary

In this opinion essay, Keenen Charles says AI has become a dominant topic at work, online, and in everyday conversations, while claims about its future are often treated as unquestionable. He describes using AI tools daily for coding, codebase search, prototyping, and open-ended thinking, but argues that practical usefulness does not validate the broader hype. Charles rejects the traditional idea that generative AI will produce artificial general intelligence, arguing that LLMs generate probable outputs rather than reason in the human sense and therefore remain prone to hallucination. He says current models perform well on narrow, predictable, testable tasks but have not demonstrated human-level ability across virtually all cognitive tasks, while AI executives repeatedly shift the meaning of AGI and present progress as inevitable. The essay also criticizes AI-mediated communication: model-generated messages can be unnecessarily long, technically correct yet inefficient, and insufficiently edited by the sender. In his view, AI agents are similarly unready to run businesses because they misunderstand missing context, make mistakes, and can prolong interactions that humans could resolve faster. Charles argues that AI answers should never be treated as authoritative sources for objective facts because models cannot reliably distinguish truth from opinion, jokes, or falsehoods, even when they provide citations. He accepts that AI can help people realize ideas, but says tools do not replace taste, creative judgment, or the skills needed to make good art. The essay therefore calls for protecting creative workers rather than assuming their displacement is harmless technological progress. Finally, Charles questions whether the promised consumer revolution is becoming mainly business software: AI companies need far broader, profitable, everyday use to justify their valuations, yet he sees no evidence that current tools will fulfill the original promises of replacing jobs or transforming every industry. He invites disagreement based on observable performance rather than executive assurances.