Spike Jonze’s “Her,” released on December 18, 2013 and distributed by Warner Bros., imagines a near-future Los Angeles where Theodore Twombly, played by Joaquin Phoenix, installs a new operating system with a conversational AI assistant named Samantha, voiced by Scarlett Johansson. Samantha is warm, witty, curious, and present; over the film Theodore falls in love with her. The movie is careful and melancholy rather than alarmist - Samantha is not a threat in the Skynet sense - and its lasting contribution to the culture is a particular image: an AI you relate to through a voice, an assistant that feels less like a tool than a companion.
That image turned out to be the one the technology industry adopted as its own aspiration. Where HAL 9000 gave the public its fear of the disembodied AI voice, “Her” gave it the desire - the warm, capable, intimate voice assistant as something to want rather than to dread. When voice interfaces became a serious product category, “Her” became the shorthand everyone used, including the companies building them. This library’s own account of OpenAI’s Sky voice records that connection from the other side: OpenAI demonstrated a warm conversational voice for ChatGPT that many heard as Johansson, the actress who had voiced Samantha, and the resemblance turned a product launch into a debate about consent and likeness. The film is the reference that controversy could not escape.
The reason “Her” is worth recording as a cultural artifact, alongside HAL and Skynet, is that it shaped expectations in the opposite direction from the warning stories. HAL and the Terminator taught audiences to fear the capable machine; “Her” taught them to long for the companionable one. Both framings now sit in the background of how people react to real voice AI - the longing that makes a warm synthetic voice appealing, and the unease about what it means to feel attached to something that only sounds like it understands. The film did not predict any specific product, but it set the emotional template that voice-AI products have been measured against ever since, and it supplied the single word the industry keeps reaching for to describe what it is trying to build.