Artificial Intelligence Should Be Banned for Its Colossal Arrogance
Is your heart racing when you think of your future with Artificial Intelligence?
Do we know, we are going forward?
Do we know, we are going backward?
Do we know much of anything in this crazy world of Deep Learning, Machine Learning, Artificial Intelligence?
The lack of preparation — does it keep you up at night?
70% of the population knows their jobs are at risk but only 30% are doing anything about it such as furthering their education or upgrading their job skills.
An excellent article is “The 4 Industrial Revolution is Here Are You Ready?” The author is Bernard Marr. He makes the point that the job market is being “segmented” into “low skill/low pay” and “high skill/high pay” segments.
CNBC has an article here: “The Future of Work Won’t be about Degrees It will be about Skills”.
“Sixty-five percent of children entering primary school will end up in jobs that don’t yet exist, reveals the World Economic Forum.” According to the CNBC article.
Another observation is that “some jobs will be obsolete.”
Adding to the turmoil, the changes (in some jobs) “might develop so swiftly, that even those who are ahead of the curve in terms of their knowledge and preparation might not be able to keep up with the ripple effects of the changes.”
The image above relates to the frog fable. That if you raise the temperature of the water too fast a frog will jump out. If you raise it slowly enough he will cook.
Mr. Marr wrote an article that appeared in the September 7, the 2018 issue of Forbes.
Massive Mining Operations at Rio Tinto Mine, Western Australia, Adobe Stock Photo
Mining produces commodities so my product is probably the same or very similar to yours.
There are no hocus-pocus retail-style tricks to pretend that my product is better than yours.
What this means is that every penny I can save in “yield, speed, efficiency can make an extraordinary impact” on profits.
The labor force for American mining and extraction is about 670,000 people.
But here is the key — almost every other industry depends on mining!
An interesting aside some of Australia’s mines have purchased Canadian mines and now an Alaskan Gold Mine is on the selling block.
“Rio Tinto’s operations include 16 mines, 1,500 km (932 miles) of railway, three ports and more.”
“The operation creates 2.4 terabytes of data every minute from all of its mobile equipment and sensors that collect data in real-time to help monitor the equipment and production.”
No doubt every piece of equipment has sensors that tell managers if a piece of equipment is not working properly and when it is approaching the time for preventative maintenance.
One new and extreme growth area for mines is “smart sorting”.
The overburden that has to be cleared to expose the precious metals is an exceedingly large volume. Using smart sorting companies can specify what they want to be sorted. “This leads to savings in fuel and energy during processing.”
A Case Study
Proctor and Gamble owns Oil of Olay. It was losing customers in droves.
It also didn’t have the data needed to push marketing thru stores like Walmart and Target.
So Oil of Olay looked “around at the number of beauty recommendation apps that were springing up and decided it could do better with the image analysis tech it had been developing for dermatologists over the last 25 years.”
Testing turned Olay into a product for a direct-to-customer website, Olay.com.
Consumers who use the tool double the conversion rate (how many sales are purchased divided by the number of customers), and when they’re buying their basket is 30% higher in sales. Those are HUGE numbers in retail.
It’s now working to shore up its products to better suit Millennials.
Millennials are more conscious of fragrances and are looking for items that are more sensitive on the skin.
All of these taken together mean that Olay has to hire people with completely different skillsets than they traditionally would have.
They have the idea and the proof of concept but they still have a long journey ahead.
A big question for them as it is for all dot-com companies is: “What to do when a customer fills the grocery cart and just leaves with everything in it?”
Another major conundrum for artificial intelligence technology is that software models tend to discriminate against certain groups like women.
Watson is the tool IBM wants to use to inspect AI-powered software to find unintentional bias when software makes decisions.
IBM made a breakthrough in the speed it takes Watson to parse models. I wrote a post on the pitfalls that befall engineers attempting to learn how to understand statistical models.
Some of the concepts students have to learn?
Statistics: Understanding Machine Learning requires solid knowledge of statistics fundamentals.
Concepts of Machine Learning Theory: This describes the understanding of the range needed to determine how different loss functions work. An understanding of backpropagation is useful, and what a computational graph is.
Data Wrangling: Applied Artificial Intelligence uses the success of a model to correlate the quality and quality of the data input. The old saying: “Garbage In-Garbage Out” true. This is especially true because of the extremely large volume of data needed for developing and testing AI models. As we saw earlier, Proctor and Gamble just didn’t have enough data to develop Olay products using statistical models.
Debugging/Tuning: Building models requires a solid knowledge of fundamentals. It is essential that the right architecture and parameters are chosen from solid theoretical fundamentals. Good infrastructure is necessary to be able to test different configurations. I wonder how many of the crashes of self-driving vehicles are due to something like this.
Software Engineering: Applied Machine Learning will allow you to leverage your Software Engineering Skills. Sometimes this task required a little twist here and a little yank there. Mastery of these skills and concepts to solid Artificial Intelligence Performance.
“2% of Americans — 7 million people lost their jobs in mass layoffs between 2004 and 2009.”
Harvard Fellow in Technology and Public Purpose Susan Winterberg shared that “laid-off workers if they are able to return to the workforce typically see a permanent reduction of 17% — 30% in wages.”
“If they aren’t able to rejoin the workforce within two months, callbacks drop dramatically no matter how good the resume and many are left to join the gig economy as contractors without benefits.
They will suffer a menagerie of woes, from depression to marital problems. Even their children suffer. They are 15% more likely to repeat a grade in school.
Mid-skill work is scarce in cities and rural areas David Autor
David Autor, MIT Professor of Economics finds that we are Importing Political Polarization.
Disenfranchised workers are rightfully angry and express it with their vote. (See my blog post on the reasons Trump got elected.)
“Both Republican and Democratic districts that have been heavily impacted by automation and outsourcing tended to oust moderate congressional representatives in favor of more conservative or liberal ones, respectively.”
Furthermore, the researchers found that “in presidential elections, counties with greater trade exposure shifted toward the Republican candidate. MAGA is the right language, offering nationalistic sentiment and relief for those caught in the technological crossfire of the inevitable future.”
But in the words of Benjamin Franklin “well done is better than well said.” In other words, workers need action.
I suggest you read my post on Trump’s Deep State if you have children. I was really impressed with the research conducted by Dr. Bobby Azarian published in “Psychology Today”; David Dunning and Justin Kruger in the “Dunning-Kruger Effect;” Sheldon Solomon in “Terror Management Theory”; and the work of UC Santa Cruz Professor Thomas Pettigrew.
Which side is going to win the coming war? The one with the best plans and policies? The one with the dirtiest tricks?
“Two Ball State researchers found that between 2000 and 2010” about 87% of manufacturing job losses stemmed from factories becoming more efficient.”
So far I have dealt with blue-collar workers. My caveat is for white-collar workers even those that are upping their skills.
George Tech has 2,350 undergraduate students in computer sciences. Master’s Students are another story. Georgia Tech is ranked 8th in the world in the number of Master of Science students that are graduated every year — 6,838 (including 6,365 online students). I will leave you to draw the conclusions.
President Obama said in his farewell address. Job losses “will come from the relentless pace of automation that makes a lot of good middle-class jobs obsolete.”
Research supports Obama’s claim. Far more jobs are lost to robots and automation (better technology) than trade with China, Mexico, or any other country.
Projections are that the U.S. job loss rate will exceed 47% — FORTY-SEVEN %!
That is in addition to the 10 million already lost by office workers, 7.5 million since 1980!
Written by Craig Martineau
Retired and Concerned for my Sons and Grandchildren