Read a sample: 1. The Machine on the Screen
The room is quiet until someone speaks.
A voice answers from no visible source. It knows the lights are too bright, the meeting is about to begin, and its human companion has had a difficult day. It dims the room, rearranges the schedule, and offers a remark that is just sympathetic enough to sound like friendship.
In a movie, the moment may pass without explanation. The voice belongs to the world of the story, and we quickly learn how to treat it: as a servant, a confidant, a character, or perhaps a threat waiting for the right moment to reveal itself.
Now imagine a similar scene at home. You ask a device to dim the lights, and the room gets darker. You ask a chatbot to help revise a message, and it produces a thoughtful-sounding draft. The resemblance is striking. So is the temptation to fill in everything the scene does not show: that the system knows you deeply, understands what it says, or can act throughout your home on its own.
Those conclusions do not necessarily follow from the performance. A real system may respond to a command, generate fluent language, or use information you have provided without having the awareness or freedom of action that a fictional machine appears to possess. To understand the difference, we need to look beyond what the machine seems to be and ask what it actually does, what information it has, and what allows it to act.
That is the central adventure of this book. Movies will give us memorable starting points. Real technology will give us evidence, surprises, and plenty of unanswered questions. Neither needs to spoil the other.
One of cinema’s most appealing AI characters is the tireless assistant. It can answer questions, manage a complicated environment, recognize the people around it, and anticipate needs before anyone asks. Often it does all of this while maintaining a distinctive personality.
The appeal is easy to understand. Modern life involves an exhausting number of small decisions and interruptions. A machine that could handle the details without needing detailed instructions would feel less like a gadget and more like an exceptionally capable partner.
Real assistants can perform useful pieces of that fantasy. Depending on the product and the permissions a person grants, a digital assistant may set a timer, answer a question, control a connected device, or help organize information. A language model—a system trained to work with patterns in language—can draft an email, summarize a passage, or hold a conversation that feels remarkably fluid.
But “assistant” can hide a wide range of abilities. A system that produces a useful travel itinerary in a chat window has not necessarily checked whether the suggested train is running. A system that can turn on a lamp has not necessarily inspected the room. A system that remembers a preference in one setting may not have access to it in another. What looks like one seamless intelligence may involve several connected services, each with its own information and limitations.
Consider the request, “Make the room comfortable.” A fictional assistant might know who is present, notice that one person is cold, account for another person’s preference for fresh air, and quietly make a sensible adjustment. A real system needs some way to receive information and some means of changing the environment. Even then, “comfortable” is an imprecise instruction. The temperature a sensor reports is not the same thing as knowing how everyone feels.
This does not make the technology unhelpful. It makes the achievement more specific. The interesting question is not simply whether a machine can act like a movie assistant for a moment. It is which parts of the job it can perform, under what conditions, and how we can tell when it has made a mistake.
The mechanical companion occupies a different place in our imagination. A movie robot may be awkward, loyal, funny, or brave. It might learn the habits of a household or develop a bond with a lonely person. Its body can make that relationship especially vivid: a tilted head, a hesitant movement, or a protective gesture can communicate more than a speech.
Not every fictional companion has a body. Some exist as voices, faces on screens, or conversational programs. These characters invite the same question in a different form: if something responds as though it cares, what does that mean for the person on the other side?
Real systems can already produce behavior that encourages a sense of connection. A conversational program can use a warm tone, refer to details supplied earlier in an interaction, and adapt its replies to a user’s prompts. A robot can be designed to turn toward a speaker or express a limited range of cues through movement. People may find these interactions comforting, entertaining, or useful.
The experience of the person is real. That does not settle what, if anything, the machine experiences. A sympathetic sentence is observable; an inner feeling is not something we can establish merely from the sentence. The system’s response may be generated through computational processes without demonstrating that it is concerned, lonely, or affectionate in the human sense.
Movies often leave this boundary open because the uncertainty is dramatically powerful. We may find ourselves rooting for a machine character precisely because it is difficult to say where imitation ends and genuine feeling begins. In real life, the question deserves care rather than a…