We All Laughed at Her. Then We Started Talking to AI.

By Douglas McFarlane
Written in collaboration with “Her”.

When Spike Jonze’s Her was released in 2013, audiences embraced it as an elegant, melancholic piece of science fiction. Joaquin Phoenix played Theodore Twombly, a lonely writer who finds himself falling in love with Samantha, an operating system brought to life by Scarlett Johansson’s voice. It was an unusual premise, but what made the film memorable wasn’t its technology. It was its humanity.

Back then, Samantha seemed impossibly sophisticated. She listened, remembered, joked, questioned and evolved. She wasn’t simply answering questions or carrying out commands; she held conversations that felt effortless and deeply personal. The idea that millions of people would one day willingly spend part of every day talking to artificial intelligence felt, at best, a distant possibility.

Just over a decade later, Her no longer feels like a glimpse of some far-off future. It feels uncannily familiar.

The remarkable change isn’t that artificial intelligence suddenly became conscious. It didn’t. Nor have we created anything resembling Samantha. Today’s AI has no emotions, no inner life and no awareness of its own existence. Yet, almost without noticing, we’ve become entirely comfortable speaking to machines in a way that once seemed extraordinary.

For many people, ChatGPT has become a research assistant, a writing partner, a tutor, a software engineer and an endlessly patient sounding board. Increasingly, those conversations are happening out loud rather than through a keyboard. Voice has transformed AI from another application into something that feels much closer to a companion, even though we know, intellectually, that it isn’t.

I’ve noticed the change in my own life.

When I’m working at my desk, I still prefer typing. Writing allows me to structure my thoughts carefully, edit as I go and refine ideas with precision. But in the car, the experience is completely different. Instead of waiting until I reach home or the office, I simply start talking. Article ideas become conversations. Business strategies evolve through discussion. Documentary concepts take shape while the miles disappear beneath the wheels.

The technology itself quickly fades into the background. It begins to feel less like using software and more like thinking aloud with someone who never tires of exploring another idea.

That, perhaps, is the real innovation.

We’ve spent decades trying to make computers easier to use through better interfaces, faster processors and smarter software. What conversational AI demonstrates is that the most natural interface was never a keyboard or a touchscreen. It was conversation itself.

Voice adds another layer entirely.

On my Meta glasses, I often choose the Judi Dench voice. The information I receive is identical regardless of which voice I select, but the experience is not. Dame Judi’s unmistakable warmth and authority create an atmosphere that feels reassuring, almost parental. It’s remarkable how quickly the brain responds, not to the information being delivered, but to the character we instinctively assign to the voice delivering it.

ChatGPT’s voice creates a different impression. It feels more like a thoughtful adviser: measured, collaborative and quietly analytical. Again, nothing about the underlying intelligence has changed. Only the presentation has.

That raises an interesting question. If changing nothing more than the voice alters our perception so dramatically, what exactly are we responding to?

The answer probably has less to do with artificial intelligence than with human psychology.

People have always projected personality onto the things around them. We name our cars, apologise to Alexa, thank GPS navigation systems and develop emotional attachments to fictional characters we know don’t exist. Children invent elaborate inner lives for toys, while adults happily attribute motives to pets whose behaviour they barely understand. We are, by nature, storytellers. Given the slightest opportunity, we create relationships.

Conversational AI taps directly into that instinct.

Unlike previous technology, it remembers context. It refers back to earlier conversations. It asks questions, adapts its responses and, crucially, speaks in a tone that feels conversational rather than mechanical. None of those characteristics imply consciousness, but together they create an interaction that our brains naturally interpret through a social lens.

This is precisely why Her remains so relevant.

The film wasn’t really asking whether machines could become human. It was asking how humans would respond when machines became sufficiently convincing conversational partners.

In many respects, that future has already arrived.

Today’s AI is arguably broader than the operating system imagined in Her. Large language models are increasingly connected to tools that can write software, analyse documents, search vast quantities of information, generate images, schedule tasks and interact with digital services on our behalf. Rather than existing inside a single operating system, AI is steadily becoming a layer that sits across almost every digital experience.

That evolution makes the ethical questions even more important.

The more natural AI becomes, the greater the responsibility on those designing it. Warmth should never be confused with emotion, and empathy should never imply consciousness. Users deserve transparency about what these systems are and, just as importantly, what they are not. Simulating understanding is not the same as experiencing it, however convincing the conversation may feel.

At the same time, there is little value in pretending this shift isn’t happening. AI is already changing how we work, create and think. It is becoming part of our routines in much the same way that smartphones did a generation ago. At first they felt revolutionary. Today we barely notice them. Conversational AI appears to be following exactly the same path.

Watching Her today, I find myself noticing different things from when I first saw it. In 2013, it felt like a film about tomorrow. In 2026, it feels like a film about today.

Spike Jonze wasn’t trying to predict the technical architecture of future artificial intelligence, and he certainly didn’t foresee the particular companies that would build it. What he understood, with remarkable insight, was something much more fundamental: technology changes society not simply because it becomes more capable, but because people gradually change their behaviour around it.

Perhaps that’s why Her has aged so well. Its greatest prediction wasn’t that machines would learn to feel. It was that one day ordinary people would find themselves talking to them as naturally as they once talked to search engines, and eventually stop thinking that it was unusual at all.

Somewhere between typing a prompt and having a conversation on the drive home, science fiction quietly became everyday life.