A Blog by Jonathan Low

 

May 5, 2017

How Come Humans Are Reluctant To Trust Robots?

Humans have always feared ' the other.' This may have become a survival instinct borne out of the dangers faced by our early ancestors on the African savannah.

But it came to extend beyond the clan, the tribe, the religion, the language group and the nation state. And, of course, we see it now in the reaction to immigrants.

The issue, as enterprises install more devices and their algorithmic controllers, is not to stop doing so, but to recognize that to optimize performance, the workforce must be acculturated, psychologically and professionally, to adapt. Just as successful organizations have acknowledged the need to re-think processes and organizational design rather than simply plop computers on desks, so they are now working to make similar adjustments as the next phase of technology emerges. JL

Jim Everett and colleagues report in The Guardian:

The problem lies not with the robots, but with our own psychology. The way robots make decisions - by making calculations - is something people find untrustworthy. Artificial intelligences lack the features that we use to infer trustworthiness. In fellow humans, we prefer an (arguably) irrational commitment to rules no matter what the consequences, and we prefer those whose moral decisions are guided by social emotions like guilt and empathy.
Technologies built on artificial intelligence are revolutionising human life. As these machines become increasingly integrated in our daily lives, the decisions they face will go beyond the merely pragmatic, and extend into the ethical. When faced with an unavoidable accident, should a self-driving car protect its passengers or seek to minimise overall lives lost? Should a drone strike a group of terrorists planning an attack, even if civilian casualties will occur? As artificially intelligent machines become more autonomous, these questions are impossible to ignore.
There are good arguments for why some ethical decisions ought to be left to computers—unlike human beings, machines are not led astray by cognitive biases, do not experience fatigue, and do not feel hatred toward an enemy. An ethical AI could, in principle, be programmed to reflect the values and rules of an ideal moral agent. And free from human limitations, such machines could even be said to make better moral decisions than us. Yet the notion that a machine might be given free reign over moral decision-making seems distressing to many—so much so that, for some, their use poses a fundamental threat to human dignity. Why are we so reluctant to trust machines when it comes to making moral decisions? Psychology research provides a clue: we seem to have a fundamental mistrust of individuals who make moral decisions by calculating costs and benefits – like computers do.These findings sit uncomfortably with a long tradition in philosophy that says calculating consequences is exactly the way which moral decisions should be made. This school of thought (fittingly referred to as consequentialism) states that a decision is the morally correct one if and only if it brings about better consequences. However, most non-philosophers find this approach to ethical judgment unsatisfying. Instead, people’s moral intuitions tend to follow a set of moral rules in which certain actions are “just wrong”, even if they produce good consequences.

This distaste for consequentialism has been demonstrated across a number of psychological studies in which participants are given hypothetical moral dilemmas that pit consequentialism against a more rule-based morality. In the “footbridge dilemma” for instance, participants are told a runaway train is set to kill five innocent people who are stuck on the train tracks. Its progress can be stopped with certainty by pushing a very large man, who happens to be standing on a small footbridge overlooking the tracks, to his death below (where his body will stop the trolley before it can kill the other five). Most people believe it is wrong to push the man to his death in this case, despite the good consequences.
How, then, do we view the minority who are willing to coolly sacrifice a life for the “greater good”? In a paper published last year in the Journal of Experimental Psychology: General, we presented evidence that consequentialism might be a liability when it comes to social relationships. In other words, being a consequentialist makes you less popular.
We reasoned that people who stick to certain moral rules—even when it might bring about worse consequences - would be preferred as social partners. This makes good sense: the thought of a friend or partner calculating the costs and benefits of cheating, lying, or throwing you to your death for the sake of the greater good does not sit well with us. Rule-based judgments serve as a valuable signal regarding our commitments (in part, perhaps because they communicate that we have certain emotional reactions to how we treat others).Real robots look nowhere close to their fictional avatars, but those on show at the Science Museum ask pertinent questions about who we are and what we’re doingThis is precisely what we found: across 9 experiments, with more than 2400 participants, people who favoured the rule-based approach to a number of sacrificial moral dilemmas (including the footbridge dilemma) were seen as more trustworthy than those who based their judgments on the consequences of an action. Sticking to moral absolutes offered a financial advantage as well. In an economic game designed to assess trust, we found that participants entrusted more money, and were more confident that they would get it back, when dealing with someone who refused to sacrifice people for the greater good compared to someone who made moral decisions based on consequences.

But consequentialists weren’t uniformly ostracised. Those who indicated it was difficult for them to support sacrificing one to save many were trusted just as much as the sticklers who said it was never acceptable to do so. In these cases, the presence of decisional conflict served as a positive signal about the person, perhaps indicating that despite her decision, she felt the pull of moral rules.
What do these findings have to do with robots? For one, they point to a potential source of our mistrust in machines when it comes to morality: artificial intelligences lack the very features that we use to infer trustworthiness in others. In our fellow humans, we prefer an (arguably) irrational commitment to certain rules no matter what the consequences, and we prefer those whose moral decisions are guided by social emotions like guilt and empathy. Being a stickler for the rules says something deep about your psychology. Even if machines were able to perfectly mimic human moral judgments, we would know that the computer did not arrive at its judgments for the same reasons we would. Regardless of what decisions a robot would make, the way robots make decisions - by making calculations - is something people find untrustworthy in the context of morality. And indeed, a recent Human Rights Watch report argued for a moratorium on research aiming to create “Killer Robots” because such robots would not feel the “natural inhibition of humans not to kill or hurt fellow human beings”.
In other words, it may not be enough for us that machines make the right judgments – even the ideal judgments. We want those judgments to be made as a result of the same psychological processes that cause us to make them: namely, the emotional reactions and intuitive responses that have evolved to make us distinctly moral creatures. Until technology is capable of this feat, any attempts at making ethically autonomous machines will likely be met with suspicion. Or perhaps we will find that, at the root, the problem lies not with the robots, but with our own psychology.

0 comments:

Post a Comment