Posts Tagged ‘bias’
Definite Moral bias behind your search results: Unless you do your due diligence to sort out the facts
Posted on: December 1, 2015
Moral bias behind your search results
Unless you do your due diligence
Remember that behind every algorithm is always a person, a person with a set of personal beliefs that no code can ever completely eradicate
What are you searching for?
Am I looking for an isolated fact? The Capital of this nation…
Or are you searching for a complex issue such as “Why is there an Israeli-Palestine conflict?”
So whenever I visit a school and talk to students, I always ask them the same thing:
Why do you Google? Why is Google the search engine of choice for you? Strangely enough, I always get the same three answers.
One, “Because it works,” which is a great answer; that’s why I Google, too.
Two, somebody will say, “I really don’t know of any alternatives.“ It’s not an equally great answer and my reply to that is usually, “Try to Google the word ‘search engine,’ you may find a couple of interesting alternatives.” And
Third, inevitably, one student will raise her or his hand and say, “With Google, I’m certain to always get the best, unbiased search result.” Certain to always get the best, unbiased search result.
Now, as a man of the humanities, albeit a digital humanities man, that just makes my skin curl, even if I, too, realize that that trust, that idea of the unbiased search result is a cornerstone in our collective love for and appreciation of Google. I will show you why that, philosophically, is almost an impossibility.
01:23 But let me first elaborate, just a little bit, on a basic principle behind each search query that we sometimes seem to forget.
So whenever you set out to Google something, start by asking yourself this: “Am I looking for an isolated fact?” What is the capital of France? What are the building blocks of a water molecule? Great — Google away. There’s not a group of scientists who are this close to proving that it’s actually London and H30. You don’t see a big conspiracy among those things. We agree, on a global scale, what the answers are to these isolated facts.
01:57 But if you complicate your question just a little bit and ask something like, “Why is there an Israeli-Palestine conflict?”
You’re not exactly looking for a singular fact anymore, you’re looking for knowledge, which is something way more complicated and delicate. And to get to knowledge, you have to bring 10 or 20 or 100 facts to the table and acknowledge them and say, “Yes, these are all true.” But because of who I am, young or old, black or white, gay or straight, I will value them differently.
And I will say, “Yes, this is true, but this is more important to me than that.” And this is where it becomes interesting, because this is where we become human. This is when we start to argue, to form society. And to really get somewhere, we need to filter all our facts here, through friends and neighbors and parents and children and coworkers and newspapers and magazines, to finally be grounded in real knowledge, which is something that a search engine is a poor help to achieve.
02:54 So, I promised you an example just to show you why it’s so hard to get to the point of true, clean, objective knowledge — as food for thought. I will conduct a couple of simple queries, search queries.
We’ll start with “Michelle Obama,” the First Lady of the United States. And we’ll click for pictures. It works really well, as you can see. It’s a perfect search result, more or less. It’s just her in the picture, not even the President.
03:26 How does this work? Quite simple. Google uses a lot of smartness to achieve this, but quite simply, they look at two things more than anything. First, what does it say in the caption under the picture on each website? Does it say “Michelle Obama” under the picture? Pretty good indication it’s actually her on there. Second, Google looks at the picture file, the name of the file as such uploaded to the website. Again, is it called “MichelleObama.jpeg”? Pretty good indication it’s not Clint Eastwood in the picture. So, you’ve got those two and you get a search result like this — almost.
Now, in 2009, Michelle Obama was the victim of a racist campaign, where people set out to insult her through her search results. There was a picture distributed widely over the Internet where her face was distorted to look like a monkey. And that picture was published all over. And people published it very, very purposefully, to get it up there in the search results.
They made sure to write “Michelle Obama” in the caption and they made sure to upload the picture as “MichelleObama.jpeg,” or the like. You get why — to manipulate the search result. And it worked, too. So when you picture-Googled for “Michelle Obama” in 2009, that distorted monkey picture showed up among the first results.
Now, the results are self-cleansing, and that’s sort of the beauty of it, because Google measures relevance every hour, every day.
However, Google didn’t settle for that this time, they just thought, “That’s racist and it’s a bad search result and we’re going to go back and clean that up manually. We are going to write some code and fix it,” which they did. And I don’t think anyone in this room thinks that was a bad idea. Me neither.
But then, a couple of years go by, and the world’s most-Googled Anders, Anders Behring Breivik, did what he did. This is July 22 in 2011, and a terrible day in Norwegian history. This man, a terrorist, blew up a couple of government buildings walking distance from where we are right now in Oslo, Norway and then he traveled to the island of Utøya and shot and killed a group of kids. Almost 80 people died that day.
And a lot of people would describe this act of terror as two steps, that he did two things: he blew up the buildings and he shot those kids. It’s not true. It was three steps. He blew up those buildings, he shot those kids, and he sat down and waited for the world to Google him. And he prepared all three steps equally well.
If there was somebody who immediately understood this, it was a Swedish web developer, a search engine optimization expert in Stockholm, named Nikke Lindqvist. He’s also a very political guy and he was right out there in social media, on his blog and Facebook. And he told everybody, “If there’s something that this guy wants right now, it’s to control the image of himself. Let’s see if we can distort that. Let’s see if we, in the civilized world, can protest against what he did through insulting him in his search results.”
And how? He told all of his readers the following, “Go out there on the Internet, find pictures of dog poop on sidewalks — find pictures of dog poop on sidewalks — publish them in your feeds, on your websites, on your blogs. Make sure to write the terrorist’s name in the caption, make sure to name the picture file “Breivik.jpeg.”
Let’s teach Google that that’s the face of the terrorist.” And it worked. Two years after that campaign against Michelle Obama, this manipulation campaign against Anders Behring Breivik worked. If you picture-Googled for him weeks after the July 22 events from Sweden, you’d see that picture of dog poop high up in the search results, as a little protest.
Strangely enough, Google didn’t intervene this time. They did not step in and manually clean those search results up. So the million-dollar question, is there anything different between these two happenings here? Is there anything different between what happened to Michelle Obama and what happened to Anders Behring Breivik? Of course not. It’s the exact same thing, yet Google intervened in one case and not in the other.
Why? Because Michelle Obama is an honorable person, that’s why, and Anders Behring Breivik is a despicable person. See what happens there? An evaluation of a person takes place and there’s only one power-player in the world with the authority to say who’s who. “We like you, we dislike you. We believe in you, we don’t believe in you. You’re right, you’re wrong. You’re true, you’re false. You’re Obama, and you’re Breivik.” That’s power if I ever saw it.
I’m asking you to remember that behind every algorithm is always a person, a person with a set of personal beliefs that no code can ever completely eradicate.
And my message goes out not only to Google, but to all believers in the faith of code around the world.
You need to identify your own personal bias. You need to understand that you are human and take responsibility accordingly.
And I say this because I believe we’ve reached a point in time when it’s absolutely imperative that we tie those bonds together again, tighter: the humanities and the technology. Tighter than ever.
And, if nothing else, to remind us that that wonderfully seductive idea of the unbiased, clean search result is, and is likely to remain, a myth.
Patsy Z and TEDxSKE shared a link.
Does Your Boss Thinks You’re Awesome? Ask him to just say you’re awesome
Posted on: January 29, 2015
Does Your Boss Thinks You’re Awesome? Ask him to just say you’re awesome
If your boss thinks you’re awesome, will that make you more awesome?
This question came to us recently, when we were working with the top 3 levels of management in a multinational.
When asked to rate their direct reports on 360 evaluations, some managers consistently rated everyone higher, and others consistently lower, than the average. We wondered if this was a result of bias, and what effect it had on the people who worked for them.
If Your Boss Thinks You’re Awesome, You Will Become More Awesome
To understand this better we looked at a larger set of 360 data to identify 50 of the company’s managers who rated their direct reports significantly more positively than everyone else on a five-point scale (that is, they gave a higher percentage of their subordinates top marks than their colleagues did, skewing the curve to the right, as in Lake Woebegone, where everyone is above average).
We also identified 31 managers who consistently rated their direct reports significantly lower than their colleagues, skewing their curves to the left.
The difference is stark: Only 18.4% of the people working for the “positive-rating” managers, or the easy graders, were judged as merely “competent” (that is, just average) compared with fully 51.4% of those working for the “negative-rating” managers, clearly the harder graders.
While neither group judged even 1% of their workers as truly problematic and in need of significant improvement, almost 14% of those working for the negative-rating managers were judged to need some improvement compared with only 3% of those working for the positive-rating bosses.
It’s hard to parse the meaning of these data.
Are the positive-rating managers indulging in grade inflation?
Do the lower ratings actually represent a more objective and deserved analysis of a subordinate’s performance? (After all, it does follow the standard bell curve.)
Or perhaps the ratings are in some way self-fulfilling, and the leaders who see the best in their people actually make them better, while those who look more critically make their subordinates worse.
We favor that second interpretation, since whether deserved or not, the psychological effect of these ratings was dramatic.
Anyone who joined us in the discussions with the subordinates of these two sets of managers would have instantly seen the impact. The people who’d received more positive ratings felt lifted up and supported. And that vote of confidence made them more optimistic about future improvement. Conversely, subordinates rated by the consistently tougher managers were confused or discouraged—often both. They felt they were not valued or trusted, and that it was impossible to succeed.
These feelings directly translated into higher or lower levels of engagement: engagement scores for those working under the negative raters averaged in the 47th percentile, whereas scores for those reporting to the positive raters averaged in the 60th percentile. This difference is statistically significant.
It’s possible that the negative-rating managers simply had more than their share of less-engaged employees, but we believe the far more likely explanation is that everyone’s engagement levels started out roughly the same and that widely different daily interactions, culminating in extremely divergent performance reviews, had a strong impact on engagement levels.
This is a particularly alarming possibility when you consider the seemingly reasonable motives of those who gave consistently lower ratings.
We frequently heard them say something like, “I want my people to get the message that I have high expectations.” Those who gave high marks to their people also had high expectations, but they were more focused on sending the message that they had confidence in their people. They truly felt that they had selected the best people for those positions, and they expected them to succeed.
And did they SUCCEED?
To see, we looked at the overall leadership ratings the two groups’ 360 evaluations. We were not surprised, by now, to see that the bosses who rated everyone lower on their performance also rated them lower on their leadership abilities, while the bosses who gave the highest marks to their teams in general gave high marks on leadership as well.
The degree of difference was startling, though—with leadership ratings averaging only in the 19th percentile for the low raters and 76% for the high raters.
And the thing is, the peers, subordinates, and other associates also rated the leadership skills of the employees working for the low-rating managers lower than those working for the high raters.
The gap was not nearly as great, as you can see in the chart below, but it was consistent and significant.
The fact that the ratings given by both the low- and high-rating managers were so different from the ratings given by others suggests that both sets of managers are biased (or that managers trying to force rank their staffs are judging them unfairly).
And it also shows that these biases and rankings have become self-fulfilling, influencing subordinates’ behavior to the extent that others ultimately can see it.
If this is so, these tough graders aren’t doing the organization any favors.
There’s an interesting study that is related to this issue called “Predicting non-marital romantic relationship dissolution: A meta-analytic synthesis.” This was a meta-analysis of 137 studies collected over 33 years with 37,761 participants.
These studies were looking at factors that cause non-married couples to break up or stay together. The number one factor that kept people together was something they called “positive illusion” – essentially that the person you’re dating thinks you’re awesome.
Is it possible, then, that if a boss thinks you’re awesome you will become more awesome? On a personal level, it’s hard to dismiss.
We’ve spoken with hundreds of leaders whose bosses thought they were awesome, we know the impact is real.
Adonis Diaries 