Using AI to give workers a “productivity score”

[Anonymous]

Just came across this article, posted just yesterday:

AI Optimizes and Personalizes Credit Risk Management to Prevent Delinquency 

It has a quote from Amyn Dhala, Vice President for AI Express within Mastercard’s Cyber and Intelligence Solutions Division, under Benefits of AI in Credit Risk Management:

"The ability to predict delinquencies before they occur prompts early action to reduce credit losses and associated collection charges. A good AI model “can detect delinquencies as early as 12 months ahead, assuming that we have the right data sources in place and [are] able to build a robust model,” noted Dhala.

[Revelate]

---

@Anonymous wrote:

Just came across this article, posted just yesterday:

AI Optimizes and Personalizes Credit Risk Management to Prevent Delinquency 

It has a quote from Amyn Dhala, Vice President for AI Express within Mastercard’s Cyber and Intelligence Solutions Division, under Benefits of AI in Credit Risk Management:

"The ability to predict delinquencies before they occur prompts early action to reduce credit losses and associated collection charges. A good AI model “can detect delinquencies as early as 12 months ahead, assuming that we have the right data sources in place and [are] able to build a robust model,” noted Dhala.

---

12 months ahead?  Huh.

[Brian_Earl_Spilner]

---

@Anonymous wrote:

Just came across this article, posted just yesterday:

AI Optimizes and Personalizes Credit Risk Management to Prevent Delinquency 

It has a quote from Amyn Dhala, Vice President for AI Express within Mastercard’s Cyber and Intelligence Solutions Division, under Benefits of AI in Credit Risk Management:

"The ability to predict delinquencies before they occur prompts early action to reduce credit losses and associated collection charges. A good AI model “can detect delinquencies as early as 12 months ahead, assuming that we have the right data sources in place and [are] able to build a robust model,” noted Dhala.

---

They've obviously never seen Minority Report.

[Ardecko]

They've probably never heard of bias in data, either.

[Anonymous]

---

@Ardecko wrote:

They've probably never heard of bias in data, either.

---

Oh they must have, and just don't care. Bias in AI is constantly mentioned in the context of ML frameworks, as in the orginal linked article from MIT Tech Review:

"Machine-learning algorithms also encode hidden bias in the data they are trained on. Such bias is even harder to expose when it’s buried inside an automated system. If these algorithms are used to assess an employee’s performance, it can be hard to appeal an unfair review or dismissal."

FICO discusses it prominently in their machine learning whitepapers.

But that startup in Boston seems particularly aggressive:

"Weir [CEO] thinks most tools on the market don’t go far enough. “Imagine you’re managing somebody and you could stand and watch them all day long, and give them recommendations on how to do their job better,” says Weir. “That's what we’re trying to do. That’s what we’ve built.”

[Anonymous]

---

@Revelate wrote:

---

"The ability to predict delinquencies before they occur prompts early action to reduce credit losses and associated collection charges. A good AI model “can detect delinquencies as early as 12 months ahead, assuming that we have the right data sources in place and [are] able to build a robust model,” noted Dhala.

---

12 months ahead?  Huh.

---

I would love to see what sort of data led to that.

I never thought income and work status would mean much in credit risk modeling when there are so many vices out there that can get a person into serious financial trouble really quickly.

I can imagine them tracking all kinds of things looking for the first signs of trouble, like the first time someone makes a charge in a casino (hotel or restaurant, not even buying chips...just being there).

Maybe people who buy junk food and cigarettes are a higher risk, based on already observed data? I have no idea without the data, but something like that might turn out to be true after analysis.

[Remedios]

I sense "Work Resilience Score" 

The D-word makes me laugh, even more so now that's it's become one of the most lucrative businesses. 

It's kinda like fortune telling at times, gather some tea leaves, call them data, throw them on the table, and "interpret" 

If you happen to be right, you're digital prophet. If you're wrong, tea leaves weren't satisfactory. 

Statistically, even a blind chicken will find enough corn to survive if enough corn is thrown inside the pen.

Does that mean blind chicken is more productive,  or does it mean it has to work longer and harder because it's not on par with its poultry colleagues who do not need to spend 8 hrs trying to do something that shouldn't take "that long" 

I have several employees at my practice who usually have 2-3 free hours each day, because they are intelligent and resourceful. They never leave any unfinished tasks. 

So, do I now throw extra work at them because they completed theirs in less time, or do I let them chill...

I opted for chill. I cannot do without them, but I can certainly do without those who are banging on keyboard for 3 hrs trying to complete a single email because everything is struggle, but based on time spent "doing things"  they would appear like star pupils. 

[Anonymous]

---

@Remedios wrote:

I sense "Work Resilience Score" 

The D-word makes me laugh, even more so now that's it's become one of the most lucrative businesses. 

It's kinda like fortune telling at times, gather some tea leaves, call them data, throw them on the table, and "interpret" 

If you happen to be right, you're digital prophet. If you're wrong, tea leaves weren't satisfactory. 

Statistically, even a blind chicken will find enough corn to survive if enough corn is thrown inside the pen.

Does that mean blind chicken is more productive,  or does it mean it has to work longer and harder because it's not on par with its poultry colleagues who do not need to spend 8 hrs trying to do something that shouldn't take "that long" 

lol @ poultry colleagues....you know I laugh out loud more from reading your posts than listening to one of the Sirius XM comedy channels, and they're pretty good sometimes.

I have several employees at my practice who usually have 2-3 free hours each day, because they are intelligent and resourceful. They never leave any unfinished tasks. 

So, do I now throw extra work at them because they completed theirs in less time, or do I let them chill...

I opted for chill. I cannot do without them, but I can certainly do without those who are banging on keyboard for 3 hrs trying to complete a single email because everything is struggle, but based on time spent "doing things"  they would appear like star pupils. 

---

Yes,  you throw extra work at them, and wring every last minute of work out of them before the buzzer goes off and they can take off their shock collars and go home.

I could never work in an analytics field that involves human behavior. Although I would love to work at FICO for a while - I'd only violate like 50-60 percent of the NDA.

[Remedios]

---

@Anonymous wrote:

---

@Remedios wrote:

I sense "Work Resilience Score" 

The D-word makes me laugh, even more so now that's it's become one of the most lucrative businesses. 

It's kinda like fortune telling at times, gather some tea leaves, call them data, throw them on the table, and "interpret" 

If you happen to be right, you're digital prophet. If you're wrong, tea leaves weren't satisfactory. 

Statistically, even a blind chicken will find enough corn to survive if enough corn is thrown inside the pen.

Does that mean blind chicken is more productive,  or does it mean it has to work longer and harder because it's not on par with its poultry colleagues who do not need to spend 8 hrs trying to do something that shouldn't take "that long" 

lol @ poultry colleagues....you know I laugh out loud more from reading your posts than listening to one of the Sirius XM comedy channels, and they're pretty good sometimes.

I have several employees at my practice who usually have 2-3 free hours each day, because they are intelligent and resourceful. They never leave any unfinished tasks. 

So, do I now throw extra work at them because they completed theirs in less time, or do I let them chill...

I opted for chill. I cannot do without them, but I can certainly do without those who are banging on keyboard for 3 hrs trying to complete a single email because everything is struggle, but based on time spent "doing things"  they would appear like star pupils. 

---

Yes,  you throw extra work at them, and wring every last minute of work out of them before the buzzer goes off and they can take off their shock collars and go home.

I could never work in an analytics field that involves human behavior. Although I would love to work at FICO for a while - I'd only violate like 50-60 percent of the NDA.

---

😂

Well, my degree is in medicine, and I understand the value of data, especially given the volume of clinical research we participate in. 

That also gives me an idea on how it can be twisted, manipulated, changed, used, abused, you name it, it can be done, but you already know that. 

Right now, the technology hasn't caught up with what you all are trying to do. 

Right idea but the means are still limited, though vastly improved from mere years ago. 

I'll wait for quantum computing to quit my whining about reliability of what humans find important,  feed it to a machine, then call it "prediction". 

It cannot predict because it didnt elect. 

I'd be satisfied with tabula rasa AI, who decides on it's own what it needs based on it's own obervations, then acts upon that.

Until then, its Humans By Proxy 

[Anonymous]

---

@Remedios wrote:

---

@Anonymous wrote:

---

@Remedios wrote:

I sense "Work Resilience Score" 

The D-word makes me laugh, even more so now that's it's become one of the most lucrative businesses. 

It's kinda like fortune telling at times, gather some tea leaves, call them data, throw them on the table, and "interpret" 

If you happen to be right, you're digital prophet. If you're wrong, tea leaves weren't satisfactory. 

Statistically, even a blind chicken will find enough corn to survive if enough corn is thrown inside the pen.

Does that mean blind chicken is more productive,  or does it mean it has to work longer and harder because it's not on par with its poultry colleagues who do not need to spend 8 hrs trying to do something that shouldn't take "that long" 

lol @ poultry colleagues....you know I laugh out loud more from reading your posts than listening to one of the Sirius XM comedy channels, and they're pretty good sometimes.

I have several employees at my practice who usually have 2-3 free hours each day, because they are intelligent and resourceful. They never leave any unfinished tasks. 

So, do I now throw extra work at them because they completed theirs in less time, or do I let them chill...

I opted for chill. I cannot do without them, but I can certainly do without those who are banging on keyboard for 3 hrs trying to complete a single email because everything is struggle, but based on time spent "doing things"  they would appear like star pupils. 

---

Yes,  you throw extra work at them, and wring every last minute of work out of them before the buzzer goes off and they can take off their shock collars and go home.

I could never work in an analytics field that involves human behavior. Although I would love to work at FICO for a while - I'd only violate like 50-60 percent of the NDA.

---

😂

Right now, the technology hasn't caught up with what you all are trying to do. 

Right idea but the means are still limited, though vastly improved from mere years ago. 

As far as Industry 4.0/machine analytics is concerned, I have to agree with that. What we call 'IIOT' (Industrial Internet of Things) is really in a constant state of flux, although I am very impressed with Microsoft's new OS for IIOT endpoints (Linux based! A penguin!) dedicated to solving some of the problems in disparate interconnectivity and most importantly, secure communication of data.

Our predictive analysis involving machines works very well because we don't select up front what data points we we think might be a predictor - we just collect as much different data as we possibly can. There are quite a lot of environmental and power sensors, for example. It's actually quite a bit more than that, and the data is streaming in constantly from factory floors. It's not like credit risk modeling, where snapshots are days/weeks apart.

After (and while) all that data is collected, it's analyzed for correlations to see if there's something there that all observed machine failures had in common. Many times we are surprised at some of the results because a piece of data that we collected didn't seem all that relevant to us - but the AI/ML software we created found it fairly quickly.