A situation we all know too well
A week ago I found myself debating the future of jobs with someone I have not met and do not know. And inevitably it reached the point of battling with information sources.
And this is where every online discussion goes off the rails.
Eventually an information source is used that is misinformation at best, fabricated information at worst, but usually just click-bait for advertising revenue.
Regardless, the article headline supported their position so it was “good enough” I suppose.
I picked up the gauntlet
This time however, I decided to pick up the gauntlet of questionable information thrown at my feet and dig into it. Give it a fair shot. And determine if there is really something in there that would make me reconsider my position.
The following is an autopsy of what I found. Like a coroner, I will use this post-mortem dissection to observe what is there, and then provide informed opinions based on what I have observed.
Because people don’t vet information
I’m doing this because people simply don’t read or vet the information they share as evidence of their opinions. They share… but they don’t consume.
And I want to highlight just how easy it is to make something look official, reinforce confirmation bias, and make a bit of money doing it.
Learn about the crime by looking at the body
The article that was presented to me was a blog post titled, ‘Automation Replaced 800,000 Workers, Then It Created 3.5 Million New Jobs‘ published on the blog The Liberty Review. I will not be posting links to these articles or sites and you will see why shortly.
That article was posted to me in a comment without any commentary or explanation. Afterall, the headline speaks for itself.
On the body you will find the wounds
What I mean is that if you spend 2 minutes evaluating an article, you will see the signs of misinformation if it exists. And it’s important to remember that every crime has a motive.
Click on the image to the right and you will see a full-size screenshot of the blog article with several sections highlighted and numbered. These are the signs of a dodgy blog.
- Google AdSense placeholder to display a banner ad. I don’t know why it did not display, but the code is clearly there.
- AdChoices product ad. I believe the parent company DoubleChoice is also owned by Google.
- Another Google AdSense placeholder. I’m not using an ad-blocker so there must be something broken in their Ad configuration. The code is clearly there.
- Banner that advertisement that links to a “free e-book”. Which of course is the top of the sales funnel to sell you something.
- Aanndd another Google AdSense placeholder. They really need to work on their ad creatives.
- Reference link to where they copied the content from.
- But really this is where the information really came from. Or did it?
- Content Network that provides low-quality blog articles just like this one with the same advertising riddled link structure.
- Banner ad for some product.
Wounds on the body
It would be easy to dismiss the blog for merely being plastered with ads and shady SEO practices. But that’s not my concern. I want to focus on the content of the post.
In particular, I want to dive deeper into #6 and #7 to see where the rabbit hole goes. To make a long story short, you will have to click two references links to get to the source, or three links if you choose the other path.
Why the discrepancy?
Because The Liberty Review site reported that they reprinted the article from Venture Beat. And they also reported the original publication of the article was on FEE. However, Entefy Blog is actually the first publication of this content that I can see. There are no more reference links to follow anyway.
Sounds a bit odd that the exact same text is published verbatim on at least four different sites. Three of them with exactly the same headline.
The path of misinformation
This is what the sequence of the sites looks like and the path of links between them.
The Liberty Review ⇒ Venture Beat ⇒ FEE ⇒ Entefy Blog
Big Red Flag
I’ll now put on my Information Coroner hat and say this article is highly suspect and likely the result of ulterior motives. Keep in mind 50% to 90% of the content is copied verbatim across all four sites. That means content is being published without any analysis or thought invested.
The Liberty Review is a click-bait site.
The content that is different is thrown in to avoid getting whacked by Google’s SEO algorithm. These articles need to rank in a Google search and be sharable for the publishers to make money on ads!
Ok, but what about that content?
This article, shared by all four sites (that I know of…), refers to a report by a very well known IT Consultancy firm that I will not name. The white-paper report is perfectly legitimate in that it serves it’s marketing purpose to inform their clients and potential clients.
I’ve read and scanned most of the report and I’m reasonably confident the information is not misleading anyone. I don’t agree with the conclusion they’ve drawn, and I’ll explain why, but that is fine.
Regarding the sources they referenced, 5 out of 12 are other reports they also published which is not ideal. But then we are looking at a white-paper used for marketing purposes.
The Smell Test
“Automation replaced 800,000 works, then it created 3.5 Million new jobs”… does that sound right? For every job you automate FOUR new jobs are created? Really?!
That declaration was actually copied verbatim out of the white-paper report. So how did the IT Consultancy arrive at that number?
Looking back at job growth
The following chart from the report shows the amount of job growth or loss by job type over the last 15 years. The jobs that could be automated were eliminated. And during that same time there was an increase of jobs that could not be automated.
Below is the statement about this chart and their conclusion.
This shift is illustrated in Figure 1, which supports Frey and Osborne’s hypotheses regarding the drivers of job creation and loss across occupations. Figure 1 shows significant net job growth has occurred in professional and management occupations, as well as in skilled trades and other associate professional and technical occupations, all considered to be at low risk of automation, At the other end of the spectrum, job losses have occurred in elementary occupations and factory work – exactly as predicted. In caring, leisure and other service occupations, where the degree of computerisation is by no means certain, the UK has seen considerable jobs growth.
The conclusion is that over the last 15 years, automation has created approximately four times as many jobs as it has lost.
Huh?! They’ve clearly skipped a few steps there.
Correlation is not Causation
Somehow they have drawn a causal line between the elimination of lower-skilled jobs with the addition of higher-skilled professional jobs. They provide no evidence or explanation of that connection unless it’s buried in one of the seven articles they’ve referenced at the end.
It’s important to remember that IT Consultancy firms are in the business of selling professional services.
Another reason for optimism is that the majority of businesses [**redacted**] surveyed recognise that technology will have a significant or very significant impact on their businesses within the next ten years. Almost three-quarters of respondents said that they intend to grow their headcount in the years to 2019.
There is no connection between the 800,000 jobs eliminated and the 3.5 Million jobs created. They simply looked at the net gain or loss across high-level categories and drew a line between them.
Will history repeat?
There is a common belief that because something has always happened, then it will continue to happen in the future. That jobs eliminated in the past by technological advances have always been replaced by new and different types of jobs. That has been true, but there is zero evidence to suggest that will continue to happen.
When 250,000 taxi cab drivers are replaced by autonomous driving cars, there won’t be a new job type that sprouts up as a result of that autonomous car. We already have auto manufacturing workers and mechanics.
That is the problem. An honest white-paper with a questionable conclusion (in my opinion), from a marketing white-paper, is then copied to created several blog posts that people share as evidence of their belief.
Confirmation bias is just a Google search away.