Matt McGee has me thinking about trust today. There’s plenty to read about trust online, what it is, and how to build it.
But can Google really algorithmically identify trust? When we look at search results, are we really looking at the pages/documents that Google trusts the most to answer our searches? Can trust be gamed?
Search people are always trying to reverse engineer search engines to understand how they make decisions about serving up results.
For the most part, these efforts focus on deciphering ranking factors and signals and then trying to increase these signals. Sometimes more naturally. Sometimes more artificially. And sometimes sinisterly.
What’s often lost in all this, are more traditional notions of trust. Merriam Webster defines trust, in part, as:
assured reliance on the character, ability, strength, or truth of someone or something.
When I think of character, I can’t help but think of concepts like morality, honesty, and fairness.
Can search engines identify these as signals? Perhaps.
After all, if someone links to the “online me” in a positive context, perhaps some modicum of trust has been implied. When someone leaves a positive review about me or my work, it seems logical that they are implicitly identifying me as trustworthy.
But don’t search engines also want to identify the untrustworthy? Don’t their users benefit from knowing of which people and businesses to be wary? In this sense, it seems that perhaps search engines, in addition to showing those that are trustworthy, may also want to identify the untrustworthy.
And then there are those with the most noble of intentions but who lack the basic competency to be deemed trustworthy.
I see this frequently in the search world. The search advisor doesn’t intend to mislead or misrepresent, but they simply don’t have the understanding to make an informed recommendation.
Can search engines algorithmically identify ability? Again, perhaps.
Context can signal ability. However, like character, quantifying ability is complex and fraught with subjectivity. For example, I could teach someone who knows nothing about chess how to move pieces on a chess board. From that person’s perspective, I have more ability than they do. However, a chess master would view my basic ability to move pieces at infantile. She would judge my ability as quite low.
And then there’s strength. Which, in the context of trustworthiness, must be relating to mental or moral strength. Which both seem to hearken back to character and ability.
Perhaps strength in this context refers to the quantity of character and ability.
Which, in terms of search signals, may be reflected by the quantitative measure of the quality and quantity of trust online trust signals.
And finally, there is truth. Which, to me, seems like the most closely related to trust. For, it seems, that the greatest measure of one’s trustworthiness, is their capacity for telling the truth.
Again, a very complex signal to ascertain from words and links around the web.
But assuming, for the sake of argument, that search engines are able to come up with proxy signals for trustworthiness, then it seems that any attempt to increase these signals, must detract from their inherent trustworthiness.
For we do not trust those that exhibit character, ability, strength, or truth for the sake of proving their trustworthiness. In fact, it is these people of who we should be the most cautious.
If this is the case, then the best we can do to build trust, is simply be trustworthy and not focus on the act of building trust.
But alas, at least at present, search engines are software, albeit sophisticated, and aren’t yet all that good at distinguishing the trustworthy from the untrustworthy. And while this isn’t a justification to act in a way that breaks online trust, it doesn’t make it any less frustrating when the untrustworthy excel at tricking search engines to believe that they are trustworthy.