Can you be fired or arrested for a crime you haven’t yet committed, but most likely will? JP Morgan thinks so – and they’ve developed new algorithms to monitor their employees and stop rogue trading behaviour before it happens.

Stay tuned to find out how and why they’re doing this.

Transcript

Amanda: Remember the movie Minority Report? It was a few years ago, and Tom Cruise’s smouldering eyes aside, the general gist of it revolved around arresting people for crimes they hadn’t yet committed, but were going to. Well, reality is becoming even stranger than fiction, as JP Morgan attempts to stop rogue trading behaviour before it happens.

And Tom Cruise doesn’t even star.

Jeremy: In a scene reminiscent of the movie Minority report, the Tom Cruise thriller, JP Morgan is using predictive analytic software to stop rogue trading behaviour before it happens. Let’s discuss.

Archana: JP Morgan has paid $36 billion in legal fees over the recent times, and since the financial crisis – and that’s led to the company unveiling a new algorithm that helps it to pinpoint rogue employees before erroneous trades happen.

Shaheerah: As Archana mentioned, they’ve paid $36 billion in legal bills. One of the famous ones was the 6.2 billion London Whale Fraud. And this both hurts their reputation and it hurts their profits as well. Now if they can actually cut legal and other expenses, the investment bank’s return on equity would actually rise to 13% as opposed to the 10% they made last year.

Jeremy: What kind of data points do you think they’re analysing to predict this rogue behaviour guys?

Shaheerah: Well, there are a lot of factors that will go into this new deduction software, some of them will include: are these workers skipping the compliance classes; are they breaking the personal trading rules; or are they reaching the Market risk limits? The whole point of this software is to predict the themes and the patterns in their employee behaviour, so it’s actually pretty cool. You’re trying to predict how a human is going to behave. And one thing I think is that they should really be keeping an eye on the middle managers and the top-level executives, as opposed to the lower level managers. It’s the top-level executives that have the access to the data, and have the authority to make the decisions, and they know all of the internal systems and controls.

Archana: Algorithmic misbehaviour prediction is apparently a booming business. Again, this all boils down to big data and data analytics which we’ve been talking about of late quite extensively. And going back to what we specialize in as well, so P4Capital we do understand the space and we do have the right kind of candidates who are experienced doing data analysis, and working with some of the latest and greatest in terms of software tools that are out there.

Jeremy: Okay, both of my colleges here have mentioned that JP Morgan has spent $36 billion in legal fees, and this will be a measure to help reduce those fees and reduce scandals that occur, but I think they’re going to be hitting some new legal actions and sanctions against them. When you try to discipline a trader for something that he hasn’t done yet – is that really going to hold up in court? You’re just opening up a whole new can of worms I think.

Archana: Absolutely, I do agree with Jeremy – it is a very tricky space to venture in, but I guess it’s more for the compliance team and the regulators to show to the institutions out there that they have the right kind of checks and balances in place, so that if JP Morgan again faces a similar situation that it has witnessed in the past, it would have its books in order.

Shaheerah: And just apart from this software, they are also spending money as well. So the company has hired 2500 compliance workers, and they spent $730 million over the past 3 years to improve operations. They’re also equating a special surveillance unit to monitor the other electronic and telephone communication in the investment bank. So if you go back and look at the fraud investigations that they’ve done, a lot of it could have been caught earlier on if they were checking the emails and the phone calls that the employees were having.

Jeremy: Yeah, a company like JP Morgan must have billions of emails going through their systems in a year. That’s a lot of work for 2,500 compliance people to be monitoring. I can see where there’s some value added to automating this to the Big Brother.

Amanda: What if people go outside of JP Morgan, and employees just send emails from their private servers?

Jeremy: That can probably be monitored as well, depending on how they’re using it. I don’t know about the legality of that.

Archana: So in the past JP Morgan and some of the other investment banks as well have used dedicated whistle-blower phone lines and email addresses where workers could actually email anonymously with tips to the management if they do see any fraudulent behaviour. Obviously this is a step up for them in terms of using technology to monitor behaviour of their employees. Also another interesting thing is that this is obviously just part one – the algorithm is just part one of the entire review of the investment bank’s work culture. The second part is obviously is HR and training related – JP Morgan is invested heavily in terms of training their employees and also to identify and fix areas where potential lapses could occur.

Shaheerah: And I also just wanted to bring some more facts to the discussion. It’s a fact that each year more than 30 million consumers actually fall victim to investment frauds – it’s a pretty big number. And the average loss for an investor is about $15,000 dollars, as well as individual losses could be all the way up to millions of dollars. And this happens because there’s nothing really suspicious in the beginning. I mean high returns and the financials seem legitimate, and this could go undetected for years. And it’s always usually in the end when you finally find out, and you start to investigate what has happened.

Archana: The new program that’s actually being rolled out would be tested in the trading part of their business first, and then the bank will spread it out throughout their global investment banking division, as well as their asset management division, and their complete roll out would happen by 2016.

Jeremy: JP Morgan has always been a market leader and they’re one of the première investment banks out there, so I imagine that there’s a bunch of other banks watching with bated breath to see if this experiment works. So only time will tell.

That was the P4Capital team discussing JP Morgan’s new initiative of stopping internal banking crimes before they happen. Want to know more? Check out our website and previous posts at www.planet4it.com or follow us @p4capital. Thanks and see you next time.

I remember going to see Minority Report when I was in high school. The movie was lacklustre, but one thing stuck out in my mind – the custom advertisements that reacted to Tom Cruise’s eyes.

For those of you who don’t remember or skipped the movie, the premise was this: advertisements (or billboards in this case) would scan Tom Cruise’s eyes as he walked by to identify him. After identification, these ads would give custom recommendations to our hero based on his data and previous purchases.

Futurist real time data at its finest.

However, headlines in recent weeks have been covering stories that seem like they could have be ripped straight from this dis-utopian world.

Real time data and facial recognition are no longer science fiction- they have become science fact.

What is it?

An individual’s face can be as unique as a fingerprint. It might even be more reliable, as a face won’t smudge or streak on surfaces.

Every face has numerous, unique features. There are the surface identifiers, like scars or skin tone, but there are also deeper, structural landmarks. These include:

Distance between the eyes
Width of the nose
Depth of the eye sockets
The shape of the cheekbones
The length of the jaw line

Most of us can recognize our friends and family by a quick glance at their faces. It happens so quickly most of us don’t even think about it. (Of course there are exceptions to this rule, but for the sake of length we’re going to skip them for now).

What Facial Recognition technology does is try to give that ability to a computer or algorithm.

Beginnings

As with most innovations we discuss on this blog, this isn’t a new idea.

It began in the mid 1960s. Scientists began to work on using computers to recognize human faces. Since those first tentative steps, the technology has come a long way.

Until recently, facial recognition software used a 2D image database to identify individuals.

Image courtesy of BBC.co.uk

For this to work, the subject needed to be looking directly at the camera. There could be little variance in light or facial expression between the database image, and the one that was being identified.

Obviously, very few pictures met this strict criteria, so facial recognition failed more often than it succeeded.

But technology never truly remains stagnant, and that’s doubly so in the digital revolution.

Facial recognition technology just made the leap from 2D to 3D. New software uses a 3D model, which allows for much looser criteria when comparing photos.

Image courtesy of Smithsonian.com

3D software uses the distinctive features of a face – the eyes, nose, and bone structure – to identify the subject.

Instead of relying on lighting or expression, this new technology relies on the structure of the face itself. It makes the software more reliable and less fallible.

Robocop?

Although these police divisions are not being led by Tom Cruise, law enforcement agencies the world over are using this technology to their advantage.

FBI

Take the FBI for example. They have just finished developing the Next Generation Identification System.

This system is now fully operational and is being used.

The FBI announced two services last fall that completed the system.

Rap Back allows officials to receive “ongoing status notifications” about reported criminal history of people “in positions of trust, such as schoolteachers.”
Interstate Photo System. This is the big one. This is a facial-recognition program that allows law-enforcement to cross-reference photographic images with criminal databases in real time.

In short, the Interstate Photo System lets the FBI use facial recognition.

Privacy groups don’t like The Next Generation Identification system for obvious reasons. They have repeated time and time again that this system ignores privacy laws.

They claim the lack of oversight raises serious civil liberty and privacy concerns.

Also, This system is largely untested. A report from 2010 found that the Interstate Photo System could potentially fail one in every five times it was used. That could lead to a lot of false positives. As it stands, that is a rate of failure higher than traditional fingerprinting – smudges and all.

Dubai

Dubai is another example of facial recognition being used by law enforcement. That Dubai is using this highly experimental technology should come as so surprise – this is a culture where the police use top of the line sports cars to keep up with the rich population.

Reports indicate they have moved one step closer to achieving real time facial recognition by adding it to Google Glass.

Did I mention that the Dubai police department will get to use Google Glass? I hope they don’t have to buy their own uniforms!

According to a Dubai Police representative, this software allows police to identify criminals’ faces and alert the detective through the glasses.

This initiative will be rolled out in two phases.

In the first phase, the technology will be used to fight traffic violations and any other vehicular offences.
In the second phase, the police Detectives will get a chance to use the wearables in their day to day crime solving business.

The New York City police department is also testing this. They began using Google Glass at the beginning of 2014, but have not yet posted the results of their tests.

Science fiction or science fact?

This technology has far reaching implications as well. It is becoming present in many different companies – not just law enforcement. Check out some of the ways that this technology is being used by businesses right now!

Mastercard

Mastercard claims it has completed a “successful” facial recognition payments trial.

Mastercard tested a beta mobile app on over 140 000 transactions. These tests involved Mastercard employees from around the world using both iOS and Android devices.

Barcelona

Every laugh at the Teatrenu comedy club in Barcelona will cost you 0.30 euros, with a price cap of 24 euros.

That’s right! Facial recognotion is being used to charger per laugh at this comedy club. The software is installed on the back of the seat in front of the customer.

The project was developed to combat falling audience numbers.

Partnering with advertising agency The Cyranos McCann, the experiment was a reaction to increased government taxes on theatre tickets, which in turn led to drops in audience numbers.

The results of the experiment have so far proved positive with ticket prices up by 6 euros, according to the theatre.

The system was so successful, it is now being copied in other theatres around Spain.

A number of people have tried, and reportedly failed, to sit through a comedy show without laughing in an attempt to get a free ticket.