Response Post-Is Data That Great?
I really enjoyed reading Kaitlyn Nastro's blog post this week entitled, "Is Data that Great?" Kaitlyn challenges the authority of algorithms as they are increasingly applied in modern life, particularly to business situations. She points out that Big Data comes down to numbers, and "sometimes numbers are wrong, sometimes numbers don't tell a story, or provide reason." In other words, using algorithms can result in recommendations that don't jive with human judgment or that result in abject error.
We cannot forget that algorithms are made by man. Which brings up another interesting point, and my reason for blogging on this topic as well. What happens when all of our latest technological breakthroughs are based on a methodology that is modeled after man's vision of the world, as well as, at least initially, our way of thinking? Do we produce something that is way smarter than us but still deeply flawed, as we are? How will we know when to let that guide us, and when to override it? With neural networks we can teach a computer to think like us. Machine Learning can analyze massive amounts of data and find the highest levels of confidence between potentially related factors, though that rarely means 100% confidence, and I would assume someone has to decide what the line between correlation and destiny is. Can you imagine letting a machine decide that, or being the source of pre-emptive decision making? What will we chose to do when that affects people's health, wealth, education and freedom? Lastly, Artificial Intelligence can literally out think us, but I assume still is modeled in our likeness.
These concepts are all based on algorithms, and the more sophisticated versions are taught to create their own algorithms, which pulls man's hand out of the equation, I think? But, do they have Conscience? Notice I'm not saying Consciousness. If you feed the entirety of humankind's philosophical thought into a machine, does it pick sides? Does it help us decide once and for all if global warming is real, that man originated from ape or god, what course of action is right from wrong, or can it only deliver a statistical probability based only on what it is exposed to, and the vantage point that that was delivered from? Does that really move the needle in how humanity answers the challenges facing future generations?
Can we feed into a machine the history of humankind, all the cultures and conflicts, accomplishments, and natural disasters, and net out at whether survival of the fittest thinking is the unavoidable over-arching reality of our species and the planet? Let's not forget that the winners have almost always written the history books, so would we even find a machine's analysis trustworthy? Can the machines bring peace? Can they prove this western-born, last man standing mentality that drives so many of our leaders is a one way ticket to extermination? Can they come up with reliable recommendations for saving the planet?
Not to get all Sci-Fi on us, but will the machines have a vested interest in our survival? Their own survival? The planet? Or, will they get selective about who among humanity are worthy? Do they care? Circling back to one of Kaitlyn's examples about how algorithms are employed to determine people's employment. A much loved teacher is fired because an algorithm measured "measurable" qualities. The answer didn't jive with view of people on the ground. What made those leaders and stakeholders go against their gut and let an algorithm's limited judgement do work that their own brains were better designed for? From my vantage point, it gave them permission to not care. Think about the risks inherent in that surrender? Kaitlyn mentions the outlying factors that make a Wall Street 'Quant's' person's algorithms fail. Aren't they only going to truly win when they can predict the unpredictable, the anomalies that go against statistical norms?
I net out with Kaitlyn on this one. Proceed with caution. Rely on our judgement above the data; it is a tool and nothing more at this point.
Looking forward to someone writing a book entitled, "Coding God" though, because that certainly sounds like where we're going.
We cannot forget that algorithms are made by man. Which brings up another interesting point, and my reason for blogging on this topic as well. What happens when all of our latest technological breakthroughs are based on a methodology that is modeled after man's vision of the world, as well as, at least initially, our way of thinking? Do we produce something that is way smarter than us but still deeply flawed, as we are? How will we know when to let that guide us, and when to override it? With neural networks we can teach a computer to think like us. Machine Learning can analyze massive amounts of data and find the highest levels of confidence between potentially related factors, though that rarely means 100% confidence, and I would assume someone has to decide what the line between correlation and destiny is. Can you imagine letting a machine decide that, or being the source of pre-emptive decision making? What will we chose to do when that affects people's health, wealth, education and freedom? Lastly, Artificial Intelligence can literally out think us, but I assume still is modeled in our likeness.
These concepts are all based on algorithms, and the more sophisticated versions are taught to create their own algorithms, which pulls man's hand out of the equation, I think? But, do they have Conscience? Notice I'm not saying Consciousness. If you feed the entirety of humankind's philosophical thought into a machine, does it pick sides? Does it help us decide once and for all if global warming is real, that man originated from ape or god, what course of action is right from wrong, or can it only deliver a statistical probability based only on what it is exposed to, and the vantage point that that was delivered from? Does that really move the needle in how humanity answers the challenges facing future generations?
Can we feed into a machine the history of humankind, all the cultures and conflicts, accomplishments, and natural disasters, and net out at whether survival of the fittest thinking is the unavoidable over-arching reality of our species and the planet? Let's not forget that the winners have almost always written the history books, so would we even find a machine's analysis trustworthy? Can the machines bring peace? Can they prove this western-born, last man standing mentality that drives so many of our leaders is a one way ticket to extermination? Can they come up with reliable recommendations for saving the planet?
Not to get all Sci-Fi on us, but will the machines have a vested interest in our survival? Their own survival? The planet? Or, will they get selective about who among humanity are worthy? Do they care? Circling back to one of Kaitlyn's examples about how algorithms are employed to determine people's employment. A much loved teacher is fired because an algorithm measured "measurable" qualities. The answer didn't jive with view of people on the ground. What made those leaders and stakeholders go against their gut and let an algorithm's limited judgement do work that their own brains were better designed for? From my vantage point, it gave them permission to not care. Think about the risks inherent in that surrender? Kaitlyn mentions the outlying factors that make a Wall Street 'Quant's' person's algorithms fail. Aren't they only going to truly win when they can predict the unpredictable, the anomalies that go against statistical norms?
I net out with Kaitlyn on this one. Proceed with caution. Rely on our judgement above the data; it is a tool and nothing more at this point.
Looking forward to someone writing a book entitled, "Coding God" though, because that certainly sounds like where we're going.
Comments
Post a Comment