How objective can we really be? Is our collective bias impacting technology?
Imagine you are driving behind another car, which suddenly starts to swerve a bit or slow down/speed up erratically.
Upon overtaking it you see a woman driver.
Nothing unusual. But what are you thinking?
Chances are you think to yourself, aha, this explains everything. Well, women with their inferior spatio-temporal skills are worse drivers then men, so this behaviour makes complete sense. You will probably experience the same reaction if you see someone struggling to parallel park.
Don't be shy, I do it too. And I am a woman!
What has just happened is that we've succumbed to the Self-Serving Bias (or Fundamental Attribution Error). This is where we look for external blame or explanation for a particular situation, usually by making a very superficial and aggressive judgement about someone else.
The subconscious objective of this (or any bias!) is that we look good to ourselves and that it cannot be our fault.
Women are in fact statistically better and safer drivers then men and have lower insurance premiums because of this. Reason aside, it's much easier to explain the situation in a binary way, rather than make an empathic assumption about the other person. Maybe a kid is choking on a biscuit in the back seat and she has to turn around causing her to drive unpredictably (this has happened to me on the motorway, it's no fun!). Maybe she needs help and not judgement.
Cognitive Biases (and fallacies) can be both entertaining and somewhat disconcerting to think about. They are a little like a slap in the face when we come across them, but they can teach us a lot about ourselves, our decision-making process and also our view of the world.
We are just human, but we are also building mini-mes that judge other humans
Biases are often thought of in the context of statistics (or sewing!) but they are alive and well in our daily life.
We all exercise and entertain them (if you deny it, you are biased about not being biased!).
We are all guilty of exaggeration, judgement, cognitive shortcuts. Because we are all human.
The challenge is to be aware enough and catch our mental slip-ups not to be blinded our righteousness (or rightness) because this can also have a negative impact on others.
In recent years biases have become a critical topic of discussion in the world of artificial intelligence in terms of ethics, equality and even democracy.
To create fair machine learning algorithms (think predictive policing, recruitment and legal decisions) the data must also be fair.
But it’s not. It’s not, because the data fed to train the algorithms is biased and unbalanced (historically) and the humans who provide the data and interpret the algorithm are also biased.
Like Garbage In-Garbage Out, we have Bias In-Bias Out. Additionally, most AI systems are so complicated and opaque that they cannot even expose the bias. Someone will just suffer.
And the people who suffer from the bias generally have little recourse of action.
Because the algorithm said so.
AI systems are already employed to evaluate teacher performance, mortgage applications, and criminal records, and they make decisions just as biased and prejudicial as the humans whose decisions they were fed. But the criteria and processes they use are deemed too commercially sensitive to be revealed, so we cannot open the black box and analyze how to adjust their biases. Those judged unfavorably by an algorithm have no means to appeal the decision or learn the reasoning behind their rejection. Many companies couldn’t ascertain their own AI’s criteria, anyway.
There is hope, however. Biased algorithms are now in the forefront of Tech Ethic discussion and many academics, professionals and legislators are demanding more accountability and transparency from the AI-driven companies.
Books such as Invisible Women: Data Bias in a World Designed for Men and Weapons of Math Destruction are also gaining huge popularity and bringing light to hairy facts that need attention and action.
But let's dive a little bit further in the wonderworld of biases and fallacies.
So what’s the difference between fallacy and bias?
Very simply, a logical fallacy is a conclusion or decision made without all information, poor reasoning, ignorance, wrong causality etc.
Spurious correlations are particularly exciting examples of fallacies.
(you can spuriously correlate as much as you like here)
Based on this chart, you can logically conclude that the more movies Nicolas Cage makes, the more people will drown (out of desperation or inspiration, who knows?). But we all know this is rubbish as a causality.
That’s a fallacy because correlation doesn't mean causality.
COVID 19 interpretation was (is) full of fallacies due to making conclusions without the full picture. This is a rather touchy topic as many people have strong opinions on either side of the believers or sceptics spectrum. Suffice to say no-one has the full picture of what is going on, leading to best possible assumptions, but also fallacies.
Cognitive biases on the other hand are beliefs, deviating from rationality in judgement and based on our own subjective reality. They are much more widely spread across demographic segments and social groups. There’s 100s of cognitive biases in various granularity and specificity. List of fallacies - Wikipedia
For example, the Hindsight bias makes us think, post-factum, that we knew something anyway. I like to call it the reverse engineering bias. It's so easy to reverse engineer an event to its constituent parts and we feel very smart by being able to derive the reason behind the event.
Your friend has had a baby girl? Oh, of course, you always knew it would be a girl because of the shape of her belly.
There was a particular political or economical crisis event? Of course you saw the 2008 crisis coming, because of all the irresponsible behaviour of the banks.
In reality we like to make sense of the facts around us and feel that we had some pre-existing expert knowledge. Most of the time we don't and accepting this fact can help us avoid unnecessary arrogance and future mistakes.
The well known Confirmation Bias, refers to a searching, massaging and interpreting of facts that support our theory. The internet is the confirmation bias' Garden of Eden. You can find information to support any argument. Often people in therapy will look for a therapist that just tells them that they are right, to support their own confirmation bias. We love “yes” people because we love to be right. But can we handle a different opinion to challenge our biases?
The IKEA effect
One of my favourite biases has to be the “IKEA effect” (or “my baby’s not ugly”). “The IKEA effect” is defined as behaviour whereby people place much more value on things that they have built themselves.
Even if it’s not straight, you flipped one of the draw liners around accidentally (and you see the raw wood - we have such a draw!), you’re still proud of it and value it more than if you would if it had been delivered already constructed (not to mention after the blood and sweat to get the damn giant packages of furniture in the car, necessitating a merry-go-round with the children seats!).
More interestingly, the researchers of this bias (in a 2011 study) discovered that people are willing to pay more for something that they would have to assemble themselves than if it’s a ready product. Go Ikea!
Apart form explaining this just as a labour of love attachment, the research also talks about “a fundamental human need for effectance – an ability to successfully produce desired outcomes in one’s environment”, where “effectance itself has multiple psychological components: actual control over outcomes and mere perceived control over outcomes “.
We made it, and therefore we feel we have a direct control of what exact shape it takes and we are in control of its destiny - where it goes, what purpose it serves and what we allow others to say about it. Therefore we become attached to it.
In a way it’s your investment in something that makes it valuable to you. Not always financial, but rather time, effort and hope.
Last last year I made my own T-shirt in a sewing course. It's definitely not Haute Couture, it does not fit well and it looks like a child made it (not the one making the H&M T-shirts), but I'm definitely keeping it. It didn't go to charity the following week and I'm wearing it with love.
That's my IKEA bias.
Could it be that Big Tech companies also have an IKEA bias of their products and don't want to acknowledge that their algorithms might actually be biased!? Who will make them see?
Comments