Jaron Lanier argues that the term “artificial intelligence” is a misleading myth, rooted in science fiction and a longing to create sentient beings. He contends that current advanced programs, like Open AI’s GPT-4, are sophisticated tools, akin to enhanced versions of Wikipedia or image search, that mash up human-created content using statistics.

Lanier believes that viewing AI as a tool, rather than a nascent mind, is more pragmatic and less likely to lead to catastrophic outcomes. While acknowledging the potential for misuse and harm, he emphasizes that human choices in design and operation are paramount. He suggests that the flexibility of these new systems, which arises from complex mathematics and vast data processing, can actually empower humans by allowing technology to adapt to individual needs, potentially ending the subservience often imposed by rigid digital designs.

He critiques the current AI safety discourse, which focuses on abstract concepts like "alignment" and "safety," as a game of outwitting a "sneaky genie." Instead, Lanier advocates for clear labeling of AI-generated content, such as deepfakes and manipulative automated interactions, along with actionable measures. This approach, he argues, fosters more intelligent management of new technologies by demystifying them and focusing on their practical applications as collaborative tools.