This article is a transcription of an episode of the RateFast podcast, which you can listen to by searching “RateFast” in iTunes or the iOS podcast store.
If you’re a workers’ compensation provider, adjuster, or case manager check out RateFast Express: the service that writes your impairment reports for you!
Do you feel secure about your job? What if you heard that a robot was about to take over the occupation that you’ve had for the last 10 years? How would you react then?
The book “Rise of the Robots” details some scenarios that may unfold — and are already unfolding — in the near future. Read ahead about the fascinating concept of robots in the workforce, and what implications it may have for society at large.
Terms
Automation (noun) – the use of largely automatic equipment in a system of manufacturing or other production process.
Artificial intelligence (noun) – the theory and development of computer systems able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
Statistical analysis (noun) – collecting and scrutinizing every data sample in a set of items from which samples can be drawn.
Interview Transcription
Claire Williams: Hello, and welcome to the California Workers’ Comp Report. Today is April 6th, and I am here with Dr. John Alchemy. Hi, John.
Dr. John Alchemy: Hey, Claire.
Claire Williams: So, today we’re going to be reviewing some relevant literature from other disciplines, and specifically looking at artificial intelligence and statistical analysis. Let’s start with this book that we read for today, called “Rise of the Robots.”
Dr. John Alchemy: Yeah, Rise of the Robots is a pretty interesting book. It was actually brought to my attention by one of our programming partners at RateFast, and it’s really an eye-opener to make people think about a couple of things: One is how automated our society has become, secondly, how quickly it’s moving, okay, because of the strength of computing power. And then finally, what is the impact to us as individuals, both from a professional standpoint, employment standpoint, future standpoint, what do we tell our kids, who are trying to figure out what to do in the future, so they’re not replaced by automation before they can even get there with their degrees? It’s just a fascinating read, I totally recommend everyone to read it.
Claire Williams: Agreed. And you heard the author on the radio a couple weeks ago. Talking about this replacement with automation, with just technology kind of coming in on the farms, machines replacing people, but those people still have jobs, because they’re able to go to the cities and find new work. But today, those jobs that were held in the cities now are also being automated. So what are we gonna tell the kids?
Dr. John Alchemy: Yeah, the book is interesting when it talks about automation, because you know, a lot of jobs went offshore to cheaper labor, and now the trend is that a lot of those jobs are reshoring or coming back to the U.S., but they’re coming back because they can be automated cheaper than the foreign labor. And what’s interesting about that is, is that again, it’s this process of the automation coming and infiltrating into all levels of jobs, and actually all services, medicine, and work comp notwithstanding. So you know, before it was like “Okay, automation’s coming in, it’s kicking out the people in the rural areas, they’re being driven into the cities,” but now, the author, when I was listening to them, this was a really nice interview with him a couple days ago on Morning Marketplace, and he’s basically saying “Look: Automation used to be sort of this thing that was very focused, like we said, with the rural situation, but now it’s ubiquitous, and now it’s more like ‘Well, what is the effect of automation on jobs,’ and it’s very equivalent to ‘Well, what is the effect of electricity on business?’” Obviously all businesses are affected by electricity, and the point being that all jobs, all industry, all economics is now being affected by automation.
Claire Williams: Right, which has a huge potential, I think, to have positive impact, but unless we address the situation of where people are going to work and what they’re going to do, we’re all going to be in a bit of trouble.
Dr. John Alchemy: Yeah, and automation is so appealing kind of to everyone because it offers accuracy, it offers things to be done more cheaply–
Claire Williams: Right.
Dr. John Alchemy: –And it has, you know, this allure of profits and all the things that go along with that. But you know, now as it’s getting here and it’s starting to impact people, A. We’re not going to be able to turn it off, and B. We now have to have a better understanding about what automation means to us, and where are we really going with all this? They’re really fascinating questions, again, we do workers’ compensation, we do analytics, we have a lot of automation in our RateFast product, but I just love to talk and think about how this automation is really infiltrating all aspects of society.
Claire Williams: Yeah, so let’s talk about kind of a Hollywood version of that infiltration with the 2011 film “Moneyball,” and kind of how it relates to statistical analysis and predictive behavior.
Dr. John Alchemy: Well, Moneyball is actually a pretty interesting, and of course, entertaining movie that came out with Brad Pitt, but the story behind it is that Moneyball was about a manager and a financial analyst, and they wanted to basically automate or objectify this process of picking players. And so the idea was that you know, the current system was based on these “experts,” they would go out, they would watch kids hit balls, they would make a determination based on their experience, who was going to be worth what, and recruited into the Major Leagues. Well, what Moneyball was all about was looking beyond that and saying “Look, there are a lot of undervalued players, and if we can apply statistics that will reflect how often they can get on base, which means how many runs we will get, we will have a new angle and a new advantage in the market.” Because they didn’t have a lot of payroll to buy all the superstars that were coming up through their team at the Oakland A’s, and then getting siphoned off a year or two later. So they set out, they looked, they came up with this statistical program, and they started to select or value players no one else could see the value in because of this system, and this automation, this approach that they would take to recruiting and making determinations on who played on their team.
Claire Williams: Mhm. And there was a lot of resistance to this method, which makes for great drama, but also as a reality, people don’t like to be replaced by math.
Dr. John Alchemy: Absolutely not. Well, people just don’t like to be replaced or questioned on their expertise.
Claire Williams: Yep.
Dr. John Alchemy: And I used to do this a lot in workers’ compensation, you know, I mean right now, there are people who will manually check the accuracy of impairment reports, and whenever they’re challenged or their expertise is challenged or their method, they become very, very defensive. And the current clients who use this method of checking impairment reports for accuracy, making determinations on values of claims, you know, they cling to this concept that the knowledge is only accessible to a precious few who claim to be the experts, and by automating the process, we’re basically demystifying, we’re finding really what is in the data set and what the value is of claims, and it’s putting at-risk these people who call themselves experts in the field. And I’m just using RateFast as an example, but this happens over and over again in some of these books that we’re citing, and definitely in Moneyball too. There’s a really great scene where all of the recruiters are at the table and you can see they’re visibly uncomfortable with the concept of removing them in favor of data management for their jobs.
Claire Williams: Precisely. And that is really well-discussed in this last book, “Super Crunchers,” which we talked about as really demonstrating the downfall of the expert when put against statistics, and the power that can come from statistical analysis.
Dr. John Alchemy: Yeah, you know, definitely for the listeners, read Rise of the Robots, read Super Crunchers, and watch Moneyball. If you can do those things, you’re actually gonna have a pretty good grasp on the concept of automation.
Claire Williams: Mhm.
Dr. John Alchemy: The undressing of the expert is very, very interesting to me personally. Again, because I am in a field that’s traditionally very conservative. Medicine, it’s very human-intensive, and comparing that to kind of education — those are the two probably remaining most human-intensive businesses, and I’m not saying that they’re automation-proof, but they are resistant, they require a lot of human input and decision-making, and that’s what makes them so costly. I mean, think about it. The cost of healthcare, constantly going up. And the cost of education, constantly going up, you know. And these are two areas of professions that are still very human-intensive. Now, when we start to automate these things, the books actually talk about the concept that radiology is probably going to be the first medical specialty to fall to automation, because it’s all about pattern recognition, and very mechanical, and less variability between interpretations. I think that people in the medical industry are fearful of this, or should be fearful of this. And it’s challenging, because we’re bringing them a system that is both helpful but threatening to them, if that makes sense.
Claire Williams: Mhm.
Dr. John Alchemy: Yeah.
Claire Williams: Yes, it’s, the question is, and definitely, this is discussed in Super Crunchers, is how to combine what can’t be automated, which he argues is this idea of human intuition, with utilizing statistics and crunching numbers.
Dr. John Alchemy: Yeah, so you know, that’s a really interesting comparison there. And I think it really comes down to is this hand-off of human intuition onto a platform that scales what the “experts” know, and then standardizes it, refines it, standardizes it, refines it, and all the time, more data being piled in. These huge databases are being compiled, and as that information comes in, the algorithms become more accurate, they become more validated, they become more consistent, and soon, it surpasses the “expert” that built it or designed it, and very quickly. Again, because the computing power is so amazing and moves so quickly.
Claire Williams: Right, so let’s go ahead and bring this right back to workers’ compensation in California and RateFast, and kind of how, how automation might really solve some of the problems we’re seeing today.
Dr. John Alchemy: Yeah, well, automation, I just want to compare, has been accepted with open arms in production for a long time. With regards to efficiencies, to product on hand, to ordering parts, having stuff set on a shelf, unsold, those kinds of things. And I always look to those industries for operations when I look for inspiration around RateFast, and what we want to do in the workspace place. For instance, right now, if a stakeholder gets a report and it’s been “reviewed by an expert,” or seen by an “expert doctor in the clinic,” and that clinic, that doctor’s written report, and serves it to the stakeholders, the problem is that the stakeholders A. don’t really know how to check the accuracy of the expert, and B., they don’t know what the context of the result is.
Claire Williams: Mhm.
Dr. John Alchemy: So, let me give you an example: I hurt my shoulder, I go see the “expert” and I’m given a value of six percent whole person impairment. Okay? And I will tell you that 99 times out of 100, most of the stakeholders are not able to pick out the data gaps in the report, calculation errors, because there are so many — measurements, what questions should have been asked that weren’t asked, and they basically look at the value, and they look at from their perspective, “Do I want what I think is a high value?” “Do I want what I think is a low value?” and so, the current system really operates on this black market intuition, you know, self-based value of the injury. And what RateFast brings in is it removes all of that subjectivity, and that anticipation of what people think the result should be. “I want a 5, I want a 19, I want a 0, I want a 6.” And we simply deliver a data set that does give a rating, okay, but the thing that it does, it gives context to the injury. So for instance, our report will provide, okay, the rating based on the algorithms and the input from the doctors, gives a result of 7 percent. But then it goes on to say, “But, this report only had 67 percent of the data present,” meaning that it was a D+ data score, okay?
Claire Williams: Mhm.
Dr. John Alchemy: So, now that stakeholders have some idea to say “Okay, I get it, this is a seven percent, but there’s only 67 percent of the data here,” then it goes on to say “Here’s what’s missing.” Okay? Here’s the missing questions, the missing physical exams, and suddenly, that expert, okay, isn’t relevant any more, because we know exactly what he or she did and exactly what’s missing. And it’s not about being, feeling like I did a good job or I’m a bad doctor or a good doctor, it’s simply about everyone saying “You know what? We’re missing five pieces of data here, this is what they are, go get them, bring them back to us.” And when you have a large database, you can validate things like that, so you can validate how much integrity of the data set is there, you can validate the accuracy of the price, you can even validate the authenticity of the data. For instance, in a very large data set that is run by algorithms and validation guidelines, you can now have a very valuable piece and a tool that you can go back and say “I have a 95 percent statistical confidence that the data set put forward in this report is correct or incorrect.” Okay? And that is a huge value, because now we have an additional piece for fraud, a very large cost on the California work comp and the DWC talk about constantly, how they want to prevent fraud, and then finally, with that rating of seven percent, the data set can be adjusted for other similar data sets, and now I can deliver a percentile of the outcome of this claim. So I can tell you now that this shoulder, missing 67 percent of the data, is a 57th percentile result based on all other adjusted shoulder injuries like it. And that is something that we have absolutely no access to right now, and that’s what’s really, really exciting in work comp, from my standpoint.
Claire Williams: Yeah, and mine too. And we think everyone should be excited about it, but there is resistance, so I’m really happy to have read these books and see, it’s resistance for good reason, I think, but I think that the solution is to really come on board as humans and say “Look, we have to partner with the robots, with the database, with the number crunching.”
Dr. John Alchemy: Yeah, I totally agree.
Claire Williams: Well, thanks so much for this interesting discussion, and we will talk to you next time.
Dr. John Alchemy: Okay Claire, thank you.
Claire Williams: Alrighty, bye-bye.
Narrator: Thank you for joining us for this episode of the California Workers’ Comp Report. You can follow RateFast on Twitter at @ratefast, or visit www.rate-fast.com to learn more.