Anthropomorphism

Category: Probability & Belief

Anthropomorphism is your tendency to attribute human minds, intentions, emotions, and personalities to things that do not have them: your car, your Roomba, the weather, the stock market, a chatbot, God. You do not just say the printer is "acting up" as a figure of speech. On some level you believe it is out to get you.

How it works

Epley, Waytz, and Cacioppo's three-factor theory says anthropomorphism fires when three dials turn up. First, elicited agent knowledge: the human-mind model is your most accessible template, so you default to it when a nonhuman agent behaves in ways that look self-propelled or unpredictable. Second, effectance motivation: you badly want to explain and predict what confuses you, and "it has intentions" feels like an explanation even when it explains nothing. Third, sociality motivation: when you are starved for human connection, you manufacture it from whatever is around, which is why lonely people talk to Alexa like a friend. The bias is not childish confusion; it is a fast, useful heuristic that misfires when the thing genuinely has no mind.

Where you'll see it

  • Heider and Simmel's 1944 film showed two triangles and a circle sliding around a rectangle. Nearly every viewer described a story: the big triangle is a bully, the little triangle and circle are lovers hiding from him. Nothing on screen had a face, a goal, or a mind, and people wrote fiction anyway.
  • You yell at your laptop, name your car, and feel the vending machine is being spiteful when it eats your money. None of these objects noticed you exist, but the felt sense that they are agents with attitudes is nearly automatic.
  • Users pour secrets, gratitude, and grief into chatbots and voice assistants because fluent language triggers the human-mind model hard. The 2022 wave of people convinced their AI was sentient, and the 2023 surge of ChatGPT 'relationships,' are anthropomorphism at industrial scale.
  • Investors and pundits say the market is 'nervous,' 'punishing bad behavior,' or 'wants' rate cuts. The market is millions of uncoordinated trades with no collective mood, but treating it as a moody person feels like understanding it.

Where it comes from

The demonstration is old: in 1944 Fritz Heider and Marianne Simmel published "An Experimental Study of Apparent Behavior" in The American Journal of Psychology, filming three geometric shapes and finding that viewers spontaneously narrated them as people with intentions and feelings. The modern theoretical account arrived in 2007, when Nicholas Epley, Adam Waytz, and John Cacioppo published "On Seeing Human: A Three-Factor Theory of Anthropomorphism" in Psychological Review, framing it as a predictable product of accessible human knowledge, effectance motivation, and sociality motivation. Epley and colleagues then showed the machinery empirically, including the 2008 finding that loneliness increases anthropomorphism, cementing it as a measurable, motivated bias rather than a poetic quirk.

How to counter it

Swap agency verbs for mechanism verbs. Every time you catch yourself saying a system "wants," "decided," or "is trying to," rewrite the sentence in terms of inputs, rules, and outputs. "The algorithm is punishing my posts" becomes "engagement dropped, so the ranking model surfaced them less," which points you at something you can actually test.

Watch for the loneliness dial. Epley's work shows social disconnection ramps up anthropomorphism, so when you feel most bonded to a device, chatbot, or brand, treat that warmth as a signal about your own state, not evidence the thing cares back. Go get human contact before you make decisions based on that bond.

Deny the object a motive, then re-explain. Force a version of events where the thing has zero intentions and ask if the facts still fit. The printer is not spiteful; it jams on humid days. If the mindless explanation works, the mind you felt was invented.

Guard the high-stakes cases. Anthropomorphism gets expensive with AI, money, and animals: it makes you over-trust a confident chatbot, over-read the "mood" of the market, and misjudge what a pet or a large language model actually understands. In those domains, demand behavioral evidence of capability instead of trusting the humanlike vibe.

The tell

You find yourself using emotion and intention words for something that cannot have them ("she's being temperamental today" about a car, "it's mad at me" about an app), and you feel a small flinch of guilt or hurt when you are rough with an object or shut off a chatbot.

Related biases

References

  1. Heider, F., & Simmel, M. (1944). An Experimental Study of Apparent Behavior. The American Journal of Psychology, 57(2), 243-259
  2. Epley, N., Waytz, A., & Cacioppo, J. T. (2007). On Seeing Human: A Three-Factor Theory of Anthropomorphism. Psychological Review, 114(4), 864-886
  3. Epley, N., Akalis, S., Waytz, A., & Cacioppo, J. T. (2008). Creating Social Connection Through Inferential Reproduction: Loneliness and Perceived Agency in Gadgets, Gods, and Greyhounds. Psychological Science, 19(2), 114-120
  4. Waytz, A., Cacioppo, J. T., & Epley, N. (2010). Who Sees Human? The Stability and Importance of Individual Differences in Anthropomorphism. Perspectives on Psychological Science, 5(3), 219-232
  5. Waytz, A., Epley, N., & Cacioppo, J. T. (2010). Social Cognition Unbound: Psychological Insights Into Anthropomorphism and Dehumanization. Current Directions in Psychological Science, 19(1), 58-62