- The Interdisciplinary Corruption Research Network (ICRN) website
- Google Podcasts
- Apple Podcasts
- Pocket Cases
- Radio Public
Tag Archives: Technology
ML for AML: Is Artificial Intelligence Up to the Task of Anti-Money Laundering Compliance?
Fighting corruption—especially grand corruption—requires effective anti-money laundering (AML) systems capable of efficiently and correctly flagging suspicious transactions. The financial institutions responsible for identifying and reporting suspicious transactions employ automated systems that identify transactions that involve certain red flags—characteristics like transaction amount, location, or deviation from a customer’s typical activity; when the automated system flags a transaction, this triggers further review. But—given the ever-increasing volume and complexity of financial transactions that occur each day, as well as the increasing sophistication of kleptocrats, criminal groups, and others in disguising their illicit activities to avoid the usual red flags—picking out the genuinely suspicious transactions can be extraordinarily difficult. Even the cleverest compliance system designer couldn’t hope to incorporate every potential red flag into the automated system.
The need to stay one step ahead of the bad actors has fueled greater interest in how new advances in data processing technology may help make automated suspicious transaction detection systems more effective. Techno-enthusiasts are particularly interested in deploying deep learning artificial intelligence (AI), as well as classic algorithms that fall under the machine learning (ML) umbrella, in the AML context. ML and AI systems extract patterns from training datasets, and “learn” (by induction) what data patterns are associated with particular identifiable categorizations. Email spam filters provide a simple example. A spam filter, which can be created to conduct a process known as classification, sorts input variables into two categories: “spam” and “not spam.” It makes its categorization based on individual characteristics of the emails (such as the sender, body text, etc.). In the AML context, the idea would be to train an algorithm with data on financial transactions, so that the system “learns” to identify suspicious transactions even in cases that might lack the usual red flags that a human designer would program into an automated system. Advocates hope that ML/AI systems could be used both to filter out the false positives (transactions which are flagged as suspicious but turn out, on review, not to raise any concerns—an estimated 99% of all flagged transactions), while also identifying unusual, potentially fraudulent behavior that may be overlooked by human regulators (false negatives). Indeed, industry experts are understandably enthusiastic about AI systems that will cut costs while improving accuracy, and proponents claim that “AI holds the keys to a more efficient and transparent AML stance[,]” urging that “[b]anks must take hold of this new [AML] weapon[.]”
To the extent that AI tools can improve upon the admittedly-clunky automated systems currently in use, it could be a step forward. But ML/AI systems have a less than stellar track record in other contexts, and a model targeted at AML compliance presents some unique challenges.
New Podcast, Featuring Franz von Weizsäcker and Niklas Kossow
A new episode of KickBack: The Global Anticorruption Podcast is now available. In this week’s episode, my collaborators Nils Köbis and Christopher Starke interview Franz von Weizsäcker (from the German Agency for International Development (GIZ)) and Niklas Kossow (form the Hertie School of Governance) about how new technologies, particular distributed ledger technology like Blockchain, can be used to curb corruption. Franz and Niklas first describe how they became interested in this topic and then, after providing a basic introduction to how distributed ledger technology works, they discuss both the opportunities and the challenges associated with deploying these new technologies to curb corruption.
You can find this episode here. You can also find both this episode and an archive of prior episodes at the following locations:
- The Interdisciplinary Corruption Research Network (ICRN) website
- Google Podcasts
- Apple Podcasts
KickBack is a collaborative effort between GAB and the ICRN. If you like it, please subscribe/follow, and tell all your friends! And if you have suggestions for voices you’d like to hear on the podcast, just send me a message and let me know.
Albanian Political Leaders Are Using Covert Tactics to Silence Anticorruption Watchdog Groups
Civil society and investigative journalism have long played key roles in exposing corruption, and many CSOs and media watchdogs—especially newer, younger organizations—now make extensive use of social media platforms to engage with the public. In Albania, for example, relatively new organizations like Nisma Thurje and Faktoje frequently expose instances of corruption via Facebook, one of the most popular social media platforms in Albania. However, corrupt politicians are taking notice of these innovative tactics and finding equally innovative ways to silence their critics. In addition to ongoing efforts to censor the media and harass activists (see, for example, here and here), the Albanian elite has undertaken more clandestine efforts to attack civil society and journalists.
One savvy scheme involves Acromax Media GmBH (Acromax), a German digital rights company owned by two Albanians, which has close ties to Albania’s ruling Socialist Party. Acromax has contracts with over 95% of Albanian television stations, with far-reaching rights to take action on its own initiative against alleged copyright violations. For example, when a civil society group like Nisma Thurje posts a story on Facebook about a politician’s corruption, and includes a link to interviews or clips of the politician’s speeches that were originally broadcast on one of those TV stations, Acromax files a complaint with the social media platform alleging a copyright violation—even though re-sharing public content that clearly displays the original source is common practice around the world and does not meet the definition of copyright infringement. Moreover, Acromax only files such complaints with respect to stories that are critical of the government; pro-government posts, including clips from these same channels, are not flagged as intellectual property infringement by Acromax.
Distressingly, even though the claims of intellectual property infringement seem bogus, Facebook has largely complied with Acromax’s demands to take down content. This may be due in part to the European Union’s recent 2019 Directive on Copyright in the Digital Single Market, Article 17 of which makes content-sharing platforms, not just individual content uploaders, liable for intellectual property violations, which in turn has caused Facebook to employ even more automation to deal with its new legal responsibilities. Unfortunately, the automated algorithms currently in use cannot reliably distinguish genuine copyright infringement from legal re-sharing, and the algorithms are sufficiently complex and opaque that it is very difficult for CSOs to challenge the take-down decisions and get their content reinstated. Acromax has exploited these weaknesses in the system to make legitimate civil society watchdogs look like serial copyright infringers. Indeed, Acromax’s harassment campaign has been so successful that two of Nisma Thurje’s founders had personal social media pages shut down because of complaints from Acromax, and Facebook further labeled Nisma Thurje “a dangerous group” and limited the range of Nisma Thurje’s social media capabilities. The technology giant further warned Nisma Thurje that its page would be shut down entirely if Facebook received even one more copyright infringement claim.
Acromax is a well-tuned operation for squelching civil society watchdogs that threaten to expose government wrongdoing, and may serve as a model for similar censorship efforts. Tackling this problem seems daunting, but these are some concrete steps that various actors—including governments, technology companies, and the civil society groups themselves—can take to address this new kind of assault. Continue reading
Leveraging Blockchain to Combat Procurement Corruption
Procurement corruption–including things like bid rigging, shadow vendors, and the steering of public contracts to politically connected firms—is an enormous worldwide problem, costing taxpayers up to $2 trillion annually. New technologies, though certainly no panacea, may offer new techniques for combating this sort of corruption. One such technology is blockchain.
Blockchain, most famous as the foundational technology for cryptocurrencies such as Bitcoin, is a “distributed ledger technology” (DLT)—a tamper-proof record of activities that are time-stamped and verified by a distributed network of computers. DLT creates a trail of information which allows for the full traceability of every transaction and stores a chronological list of transactions in an encrypted ledger. Transactions are bundled into a secure and identifiable block and then added to a corresponding chain. The blockchain is maintained and verified by the distributed crowd, eliminating the need for hierarchy and any centralized authority or middleman. And while blockchain is best known for its role in making cryptocurrencies feasible, it also has a range of other applications, including anticorruption applications. For example, Tanzania has utilized the technology to weed out “ghost workers” from the public sector, ending the monthly outflow of 430 billion Tanzanian shillings (approximately US$195.4 million) in salaries to fake employees who exist only on paper. Nigeria’s customs service has also used blockchain technology to store information on financial transactions and share these transactions across multiple computer networks.
Blockchain technology could also be used to combat common forms of procurement corruption, particularly those that involve after-the-fact tampering with submitted bids and supporting documentation. Such a system would work as follows: Continue reading
A Low-Cost, No-Tech Solution to Petty Corruption: Stickers
On a recent trip to Myanmar, I was surprised to find that all of my dining receipts came with government stickers on them. It turns out that these stickers are a solution to a tax fraud problem. Restaurants are supposed to charge sales tax, but the government has limited capacity to ensure that the tax collected from patrons actually reaches the government. So the government sells stickers to restaurants that say the price the restaurant paid, and restaurants post these stickers on each receipt for the amount of the tax. Compliance is secured through a combination of direct enforcement and public pressure. This low-cost, low-tech solution ensures the flow of money to the government instead of the pockets of unscrupulous business owners.
The same innovation could be applied to combat petty corruption, helping to ensure that the money from various charges paid by citizens—from license fees to road tolls to other government service charges—flows to official coffers rather than bureaucrats’ pockets. In any situation where an individual has to pay the government – from garbage collection to healthcare to speeding tickets – demanding a stickered receipt could ensure that the government agent doesn’t pocket some (or all) of the payment. Moreover, using these stickers would have a more subtle secondary benefit: fixing the price of government services. Consider a citizen who applies for a driver’s license and has to pay a cash fee. The bureaucrat in charge of processing the application not only has an incentive to not only pocket the cash, but also to exaggerate the size of the license fee in order to have more to steal. The stickers help ameliorate this problem, because a citizen who demands proof of payment in the form of stickers diminishes the incentives of the bureaucrat to inflate the price.
Of course, while the sticker system helps address petty embezzlement, it does not (directly) address the problem of petty bribery. A bureaucrat could demand an additional bribe on top of the official price for a service. Or a government agent could offer to not impose some charge or fine in exchange for a bribe paid directly to the official. The classic example here would be a police officer offering to look the other way on a traffic offense in exchange for a payment. Nonetheless, the sticker system may also help to curb these sorts of petty bribery, for a few reasons: Continue reading
Can Blockchain Help Bypass the Problem of Corruption in Development Aid?
Corruption undermines the effectiveness of foreign aid. While precise numbers are hard to come by, numerous press reports suggest that mass “leakages” (a euphemism for probable theft) are all too common. UN Secretary General Ban Ki Moon has reportedly asserted that approximately 30% of foreign aid is lost to corruption, though controversy over the magnitude and impact of the problem remains (see, for example, here, here, and here). The perception of a severe problem has naturally led to searches for innovative solutions, including technological solutions. One possibility that has been garnering some recent attention is blockchain technology. In fact, a few months ago, the Ministry of Foreign Affairs of Denmark, the think tank Sustainia, and the blockchain currency platform Coinify jointly published a report delineating how blockchain technology can be used to “hack the future of development aid.”
Blockchain systems make use of a shared digital “ledger,” in which each transaction contains the history of all previous transactions; because the ledger is transparent and distributed across many computers, rather than stored in a centralized database, it is (allegedly) not susceptible to manipulation or hacking, and ensures the transparency of all transactions (though not necessarily the real-world identities of those engaged in those transactions). Blockchain is probably best known as the technology that makes possible Bitcoin and other so-called cryptocurrencies. But blockchain technology and its applications are rapidly evolving, and many have already begun to see how this technology can be used as a tool to combat corruption, for example by increasing transparency in land records and by using blockchain systems to support anti-money laundering efforts. Now, companies such as Disberse, AID: Tech, and Donorcoin are developing blockchain-based fund management systems that, their proponents contend, can help reduce corruption in development aid. Blockchain technology would allow donors to transfer money to end users directly (and instantaneously), bypassing the formal financial institutions and corrupt bureaucracies that have often been the source of financial leakage, and preserving a transparent record of all transactions. This would help ensure that aid money goes to where it is intended to go.
National Digital Currencies Raise New Risks of Grand Corruption
In 2017, you may have heard of this thing called blockchain. The technology, which works by creating a decentralized, encrypted, and independently verifiable ledger of transactions distributed over a network of computer systems, has allowed innovations in the design of secure systems for recording votes, registering land ownership, and confirming digital identity. The most famous application of the blockchain, however, has been the creation of digital currencies such as Bitcoin, Ethereum, and Ripple. Many private individuals consider these currencies to be the way of the future, and the death knell of the central banker: universal, transparent, and valued according to mathematical laws rather than political preferences, cryptocurrencies—according to their proponents—will bring with them immeasurable benefits, among them making the fight against corruption easier by allowing all interested parties to “see the entirety of any transaction instantly and accurately.”
But private citizens aren’t the only ones who have heard of the blockchain: the same central bankers who are meant to be rendered irrelevant by the advent of cryptocurrencies have also taken notice. Several governments, including those of Israel, Russia, China, Estonia, Sweden, and Venezuela, have announced plans to create their own national digital currencies (NDCs) based on blockchain technology. While there are several sound economic reasons for introducing an NDC, governments frequently cite the same anticorruption benefits mentioned above.
However, there are crucial differences between NDCs and cryptocurrencies like Bitcoin. Rather than open architectures enabling full financial transparency, most NDCs currently plan to use some form of centralized ledger, giving government authorities (and only them) the ability to see and police transactions. While such centralized transparency will give honest governments a much-needed boost in the fight against corruption, it will also give oppressive and kleptocratic regimes another tool with which to steal from and oppress their populations. Continue reading
Legalized Sports Betting in the United States: Analyzing the Impact of Legalization on Corruption Risk
The rise of corruption in sport has captured the attention of many anticorruption groups, including Transparency International and the United Nations Office on Drugs and Crime. Sports corruption takes many forms, but one of the most prevalent is match fixing, which occurs when players or officials alter the outcome of a sporting event in a way that benefits those who bet money on those “fixed” games.
In the United States, concerns about match fixing, among other things, led Congress to enact the Professional and Amateur Sports Protection Act (PASPA) in 1992. The Act prohibits most states from legalizing sports gambling, with only Nevada allowed to offer betting on single games. Yet PASPA failed to curb gambling on sports, mainly because bettors turned to the black market; each year, Americans gamble an estimated $150 billion-$400 billion in illegal sports betting.
PASPA appears to be in legal jeopardy: Last December, the U.S. Supreme Court heard oral arguments in the case of Christie v. National Collegiate Athletic Association (NCAA), and while a decision in the case is not expected until later this year, legal experts believe that the Supreme Court will invalidate PASPA. This would provide all 50 states with the opportunity to legalize and regulate sports betting in their state. With that in mind, it is important to consider the effects that legalized sports gambling may have on bribery in professional sports.
Reforming the ANDSF Payroll Management System: The Limits of a Technological Solution
In my last post, I discussed the how the problem of “ghost soldiers”—soldiers who are inaccurately listed as on active duty, for purposes of generating salary payments that are then stolen—adversely affects the capacity and readiness of the Afghan National Defense and Security Forces (ANDSF). To make things worse, not only is the government making salary payments to soldiers who don’t exist, but some ANDSF personnel who do exist are not receiving the full salaries they are due. Approximately 20% of Afghan National Police (ANP) and 5% of Afghan National Army (ANA) personnel are paid in cash through so-called “trusted agents,” who are supposed to facilitate salary payments to ANDSF personnel when electronic funds transfers (EFTs) are not possible, but according to reports, corruption in the system could take as much as half of an employee’s salary. And while most ANDSF personnel receive their salaries via EFT to their personal bank accounts, this only reduces the threat of pilfering in the final distribution stage; it does nothing to correct for errors, either intentional or inadvertent, generated earlier in the process.
What can be done about these problems? The U.S.-led multinational military organization working with the Afghan government to reform and strengthen the ANDSF, known as the Combined Security Transition Command-Afghanistan (CSTC-A), is applying a technological band-aid that focuses on implementing a set of computerized systems that track personnel and pay. While these measures are helpful, they do not fundamentally change the incentive structures that drive corruption, and so are unlikely to represent a long-term solution, particularly after direct U.S. involvement winds down.