Is there a cognitive bias for ascribing agency to inanimate objects?

Question

I am curious to know if any research might have been done on this topic or if it is any way even described in scientific literature.

It's not uncommon for people to ascribe agency to inanimate objects and when objects "act" in ways that are contrary to the person's expectation (a computer that is loading files slowly, a car that won't start, an appliance that breaks frequently) some people can become angered and feel that this specific object is behaving maliciously towards them in particular.

Even when we rationally know that a computer program does not have consciousness and cannot intentionally target us we can still feel persecuted by it. I know there are a few cognitive biases that deal with incorrectly attributing malicious intent to other people (attribution bias, hostile attribution bias, etc) but I have not heard of one regarding inanimate objects and yet it clearly exists. As a software developer I find this interesting as it impacts user experience. As an example, if software attempts to be overly personable by calling a person by their name and interacting with them so as to give the illusion that the software is "alive" might it increase the potential for people to ascribe malice to it when it behaves in ways they feel it should not or when there are frequent errors.

I want to be clear I am not suggesting that people literally or consciously believe that inanimate objects are sentient and possess agency, but that is how they behave and likely what at some level increases their frustration. When a person, for example, takes a tool that has been on the blink for a while and out of frustration smashes it or tears it apart I do not find it unreasonable to believe that the relief they feel is a shot of dopamine that they are getting for "justly" punishing the tool for it's "behavior". How many times have we heard a child or maybe even an adult exclaim, "That thing hates me!" Rationally they know the thing in question is not conscious yet at some level they feel it is actively persecuting them.

A classic example is in a Canticle for Lebowitz when one of the characters is having issues with a computer that is constantly misbehaving exclaims,

"That thing-” he waved irritably toward the Abominable Autoscribe- "is a damned infidel or worse."

I was asked for scientific evidence, but I cannot provide any because that is in fact what I am asking about. Is this recognized as a cognitive bias and if so, has it ever been studied?

Personal information in messages (UX StackExchange)

Malicious Objects, Anger Management, and the Question of Modern Literature

Evil by Design: Interaction Design to Lead us into Temptation

Answer 1

Fortunately, some psychologists have looked into it, and in accordance to your observations, people do tend to attribute intentional agency to inanimate objects.

In the seminal work by Heider and Simmel (1944), participants were shown shapes of a triangle and disc moving around a two-dimentional landscape. They were not moving in random directions, but in ways that convinced the participants they were acting, fighting, playing or helping each other. When objects move in a purposeful manner, and not just fall to the ground aimlessly, we naturally tend to think they have intentions, belonging to the kind of beings with some kind of intelligence. See also Barrett and Johnson (2003): The Role of Control in Attributing Intentional Agency to Inanimate Objects.

Further more, this tendency is used by the camp of the skeptics, like Dan Dennet or Steven Pinker, to explain why people recognize many events around them as the actions of of gods or super-natural beings. By coincidence, they see some random event or movement, which has no obvious or natural explanation. The witness of such event looks for the agent behind it, and when one can't be found, it is attributed to a hidden or unseen intentional being. The way from there to gods, deities and religion, these thinkers claim, is very short. See Dennett's BREAKING THE SPELL.

Answer 2

Here is an interesting article on NPR about an AI version of the famous Milgram experiment (i.e. the psychological study of obedience to authority in which people were asked to give confederate researchers increasing shocks for making errors, http://en.wikipedia.org/wiki/Milgram_experiment). The interesting finding here is that the study was replicated using robots that had to have their operating systems/memories destroyed - participants had a harder time doing this after spending time with a helpful, kind robot.

http://www.npr.org/blogs/health/2013/01/28/170272582/do-we-treat-our-gadgets-like-they-re-human

Lots of other good references in the article above.

Answer 3

Computers interact with you and so we treat them like other, real people. Even expert MIT computer users behave this way. We all do it. I'll try to find the research on it. Someone actually wrote a book about this.

We also tend to blame ourselves when we can't accomplish something on the computer. This is great for computer programmers, but bad for people. The truth is a lot of error is attributed to bad design. So rather than feeling stupid, you can put the locus of control outside of yourself and say whoever made this program didn't empathize with you enough and perform enough tests or user observation.