Issues in Financial Crime, Technology, Artificial Intelligence and Machine Learning

By Christine Duhaime | March 5th, 2017

3rd Annual Vancouver FinTech Conference

The 3rd Annual Financial Technology Conference in Vancouver takes place on May 3, 2017, which we co-sponsor. This year’s Conference focuses on “FinTech and Financial Crime” and will cover the nexus between technology solutions to address regulation (RegTech), law, money laundering, terrorist financing, compliance, fraud, identity management, transnational criminal organizations, corruption, real estate tech, on-boarding / de-risking, predictive AI and cyber-security.

The idea is to advance the dialogue to deep-dive into the issues at the level where it matters – law, policy and financial transactional monitoring and explore solutions that work, including importantly, artificial intelligence and financial crime.

In the lead up to the conference, we’ll be populating this page with those topics to identify some of the issues worth exploring that the experts at the Conference can address.

1. Issues in cyber-security

Cybercrime costs the global economy $500 billion annually and ransomware attacks rose to 4,000 per day in 2016, forcing companies to pay $400 million in ransom in Bitcoin in 2016. As the world becomes connected (20 billion devices by 2020) with robots, cars, medical equipment, phones and refrigerators all inter-connected, the threats will grow.

Americans have said plainly that we are “in the fight for our digital lives and are not winning the cyber-security war” from China to Russia, hackers are stealing intellectual property from tech firms on a daily basis, as well as financial data and health care records. In addition, Americans have said that terrorists are crowd sourcing the murder of innocent people.

Cyber criminals are winning for many reasons including that there are more criminals than law enforcement can deal with; the law has not kept up with cyber attacks and nor is there the budget to address it properly; there are serious information challenges between the public and private sectors because the private sector does not report cyber crime; there is no deterrence in the cyber realm to prevent further cyber attacks; and terrorists are using social media and end-to-end encryption on their phones to cover their tracks.

The US is also noting that critical infrastructure is a critical vulnerability with criminals now leaving their digital fingerprints on purpose as a warning.

The battlespace is –> your cellular phone.

2. Predictive AI & data vacuuming in crime

One day soon, a machine will identify money laundering activity all on its own and will alert law enforcement on how to locate the person with the evidence all nicely packaged for prosecution. Already, in the US, machine learning analyses massive amounts of data for the SEC to find insider trading and machine learning takes recordings of faces, plugs them through facial recognition and informs police when the same people interact on street corners (druggies), visit a night club frequently (druggies), a high end hotel (prostitution) or a parking garage (car thieves).

Using data vacuuming, governments and law enforcement are able use machine learning to predict criminality. ASA – automated suspicion algorithms convert inputs into outputs. For example, an Instagram account as an input uses a person’s own data and their behavior and coverts it into predictions on the likelihood of criminality and do so in a way that is automated to cut down on the time a human would take to identity the data and come to the same conclusion. The advantage of machine learning data that was vacuumed from an Instagram account is that it can pick up correlations and connections to other people and repeat patters of behavior that a human cannot detect without years of analysis.

Accessing big data and data vacuuming and using machine learning with that data may be viewed as a form of constructive knowledge, legally speaking, allowing for law enforcement to have constructive awareness of predictions of criminality, ignoring the individual reasonable suspicion requirements. One scholar suggests that the constructive knowledge doctrine would turn our police agencies into “something like Star Trek’s Borg Collective” in that police could rely on databases of what all law enforcement agencies know everywhere in respect of a person suspected of a crime.

In anti-money laundering law, since it is criminal law, the money laundering component requires having reasonable grounds to suspect that certain conduct is tied to a criminal offence and that is a  legal determination that our courts have ruled cannot be “technical”. So then, how do we design machine learning technology for anti-money laundering detection that is, by its design, technical, that reasonably matches what a lawyer, judge or law enforcement officer would determine is reasonable to ensure it is constitutional? No one yet knows. For example, in the example above, the person who interacts at the street corner frequently could be a girl selling Girl Guide cookies and not a drug trafficker.

For an ASA to be sufficient to justify reasonable grounds to suspect and justify a search or seizure or reporting, it must be accurate and be based on an understanding of the law. ASA is only as good as its coding and if that coding is not done by lawyers, it will fail on the constitutional threshold.

3. Money laundering tech

Some recurring issues affecting money laundering is the lack of manpower to investigate cases. Prosecutors have found that when tips come in from the public (informants or whistle blowers) about money laundering or transnational criminal organizations, the information tends to be accurate and point law enforcement immediately to the key players and where the crimes can be detected. Those cases tend to be the ones that are pursued by law enforcement.

What type of money laundering tech is being developed, and does it include mechanisms to allow the public to report money laundering cases?

4. China and money laundering / corruption

China was slow to implement anti-money laundering controls and its regulatory effectiveness is “relatively low because [for ten years], a lot of money laundering activities were not detected and there was much capital flight from China” (Shan Xi Normal University).

From 2008 – 2013, Chinese financial institutions filed 45,000,000 suspicious activity reports annually to their FIU. Yet although it files 187 times more suspicious activity reports than the US, it averages about 7 arrests a year for money laundering.

The relevant law in China for money laundering that addresses the predicate offenses have some limitations compared to other countries. For example, the predicate offenses are drugs, gangsterism, smuggling, corruption, bribery, destruction of financial records and financial fraud, gambling and tax crimes.

The percentage of predicate offenses for money laundering in China:

  • Corruption 42.08%
  • Financial fraud 11.76%
  • Destruction of records 8.6%
  • Drugs 8%
  • Smuggling 8%
  • Tax crimes 5.88%
  • Gambling 4.98%
  • Gangsterim 2.71%

Types of customers involved in money laundering in China:

  • Politically exposed persons 57.58%
  • Private companies and their shareholders 27.7%
  • Cash management companies such as insurance, real estate and securities firms 15.15%

A report from a university on money laundering refers to Canada as a jurisdiction where politically exposed persons launder money and “abscond overseas.” Given that China has lax money laundering compliance itself, what does that mean for Canada, as a key destination?

5. De-risking

The past 15 years has been unprecedented enforcement action in connection with anti-money laundering law such that in 2014, over $15 billion in fines were levied by US regulators, causing de-risking as banks attempted to limit their risks. However, there is concern over the chilling effect this has on cross-border trade because banks are eliminating correspondent banking relationships. The fear of liability and of AML / CTF compliance is killing banks in high risk areas. According to two studies, between 31% to 60% of banks say AML / CTF costs are the reason for de-risking of correspondent banks. Other surveys 40% of banks and corporations said that AML / CTF was a significant impediment to trade finance and over 70% said they declined transactions because of AML / CTF.

Topics coming: 

6. Transnational criminal organizations

7. Real estate in Vancouver, Toronto, Miami, New York

8. Identity management

9. Terrorist financing

10. Law (RegTech)

11. Politically exposed persons

Share this Post:
  • Facebook
  • Twitter
  • LinkedIn
  • Print
  • email

Comments are closed.