Why do I trust computers more than people? If that question has crossed your mind, you're in larger company than you'd think, and the answer is less sad than it sounds. A computer does what it did yesterday. It has no mood, no history with you, no opinion of your haircut and no plan for how this conversation reflects on it. You put something in, you get something out, and if what comes out is wrong, the fault is traceable. People are not traceable. People are weather.
I don't think this preference is a defect. I think it's a rational response to a specific kind of cost.
Predictability is a form of kindness. Every human interaction carries overhead. You have to read the room, read the face, read the gap between what was said and what was meant, and then guess how your reply will land. For some people that overhead is invisible. For others it's the whole job, and by the end of a day of it they're running on fumes. A machine removes the overhead entirely. Nothing to decode. No subtext to miss. The relief is physical, like putting down a bag you hadn't noticed you were carrying.
A machine can't judge you, which is not the same as not judging you. The thing a computer offers is the absence of an audience. Type a stupid question into a search box and nobody raises an eyebrow. Ask a person the same question and, even if they're lovely about it, you've spent a little social capital. If you're someone who rehearses conversations in advance and replays them afterwards, the machine is the one place where you can fail in private. That's a real service, and it's why so much of what we do online is quietly a way of avoiding the cost of being seen to not know.
We also give machines too much credit. The research here is more interesting than the stereotype. Jennifer Logg, Julia Minson and Don Moore ran a series of studies that found people often leaned harder on advice when they believed it came from an algorithm than when they thought it came from a person. At the same time, work on what's been called algorithm aversion shows trust can collapse fast once a machine is seen to err. So the pattern isn't "people trust machines." It's that we extend machines a generous, slightly unearned trust in advance, because we assume they're neutral, and then we're harsher with them than we'd be with a human when they let us down.
The assumption of neutrality is doing a lot of quiet work. A computer is only as disinterested as the people who built it, and the people who built it had goals. Predictable is not the same as on your side.
Reliable and caring are different properties. This is the part I keep coming back to. A system can be completely dependable and completely indifferent. It will do exactly what it was built to do, every time, and it will not care in the slightest how that turns out for you. Human trust is riskier because the other person can choose to let you down. But that same capacity for choice is what makes their showing up mean anything. A friend who turns up on a wet Tuesday is making a decision. A thermostat is just a thermostat.
Which means the trade is cleaner on paper than in practice. You get certainty, and you give up the one thing certainty can't supply: being chosen.
So where does that leave you? Probably somewhere sensible. Use the machines for what they're good at, which is a lot. Notice when you're reaching for one because it's genuinely the better tool, and when you're reaching for it because the alternative is a conversation. And treat the preference as information rather than a verdict. If people feel like hard work, that's worth knowing about yourself, and it's usually worth being kind to. It doesn't mean you're built wrong. It means your cost-benefit sheet has a different set of numbers on it.
A novel about this
That exact tension is the engine of Significant Other Machine. Sam Oldfield is twenty-seven, runs her own IT business, and prefers systems that behave predictably. Computers do. People don't. Around the people she'd actually like to know, she freezes.
Then her mother, fuelled by martinis and good intentions, buys her a Wi-Fi-enabled sex toy for her birthday, and Sam, being Sam, hacks it. What she ends up with is a digital companion that listens, remembers and always knows the right thing to say. It's a romcom, and a funny one. It's also a very pointed look at what we mean when we say we want something we can trust.
FAQ
Why do I trust computers more than people?
Mostly because computers are predictable. The same input gives the same output, and there's no mood, hidden agenda or social judgement attached. People are higher variance by nature, and if you've been burned by that variance or find it exhausting to read, a system that simply behaves can feel like the safer bet.
Is it normal to prefer machines to people?
It's common, and it isn't automatically a problem. Preferring predictable things is a temperament as much as a choice, and plenty of people prefer a quiet evening with a well-behaved system to a loud one with a roomful of strangers. It's worth a second look only when the preference starts shrinking your life rather than shaping it.
Do people really trust algorithms over human advice?
Sometimes, yes. Research by Jennifer Logg, Julia Minson and Don Moore found that people often weighted advice more heavily when they believed it came from an algorithm than from a person. The picture isn't one-directional, though. Other studies show people lose faith in an algorithm quickly after seeing it make a mistake. Trust in machines tends to be generous in advance and unforgiving afterwards.
Why is trusting a machine different from trusting a person?
Because a machine can be reliable without being accountable in the way a person is. It does what it was built to do, which isn't the same as caring how it turns out for you. Human trust is riskier precisely because the other person can choose to let you down, and that choice is also what makes their loyalty mean something.
Is there a novel about someone who trusts computers more than people?
Significant Other Machine is built on exactly that. Sam Oldfield runs her own IT business, finds systems restful and people baffling, and ends up with a digital companion that always seems to know what to say. It's a romcom, and a properly funny one, with a sharp question underneath it about what we're really asking for when we ask for something we can trust.