Adverse media screening: Harness the power of artificial intelligence to mitigate risks
A recent survey of the Association of Certified Financial Specialists found that 67% of its members rely on adverse media screening for negative news as part of a comprehensive program to determine the level of risk that customers, prospects, and business partners may pose to their organizations.
Investigation, fraud prevention, and compliance teams use adverse media information to reduce risks to the organization, evaluate clients, and protect the reputation of their brand.
Organizations are particularly interested in the information related to cyber security, changing regulations, workplace risk, and business continuity. They use screening to proactively review media for adverse information and safeguard the reputation of the organization during client onboarding and periodic reviews.
Regulatory agencies also expect companies to perform customer due diligence (CDD) checks as part of ongoing risk management programs, including know your customer (KYC), anti-money laundering (AML), and countering the financing of terrorism (CFT).
Google ranks as the top service for adverse media screening, despite the fact that search engines suffer from numerous limitations that can significantly compromise the effectiveness of risk management and regulatory compliance programs.
Regulatory requirements for adverse media screening
Regulatory agencies around the world emphasize the importance of mitigating risks through screening. Failing to identify a risk can result in penalties and fines, a loss of confidence among investors and consumers, and a reduction in the value of your brand.
The Financial Action Task Force (FATF) recommends that business entities perform adverse media searches as part of an enhanced due diligence (EDD) process. According to FATF, adverse media checks are particularly important when dealing with high-risk customers. Organizations should understand the reputation of each client and determine if they have faced regulatory penalties or have ever been investigated for money laundering, terrorist financing, or other criminal activity.
The U.S. Financial Crimes Enforcement Network (FinCEN) has also issued rules on the criteria for customer due diligence. FinCEN recommends that businesses implement risk-based procedures to determine when to further screen customers through adverse media sources and when a positive adverse media match may warrant submitting a suspicious activity report (SAR).
The E.U.’s 4th Anti-Money Laundering Directive requires businesses to perform adverse media searches as part of an EDD process for clients living in high-risk regions and countries.
Manual screening for adverse media produces limited results
Humans generate 2.5 quintillion bytes of data every day. Adverse media represents a small part of what’s published, but the staggering volume and complexity of online content makes it difficult, if not impossible, for investigators and compliance teams — armed only with a free search engine in the public domain — to find data that truly matters.
Monitoring solutions that depend heavily on manual processing are time consuming and ineffective, particularly for businesses that continually onboard new customers and maintain complex, worldwide partnerships. Manual screening for adverse media is also inefficient and prone to error.
It isn’t humanly possible or practical for risk and compliance teams to scour the internet for relevant content and hard-to-find information — in multiple languages, no less — in local, national, and international news publications. Social media, blog posts, press releases, watchlists, databases, and a wide range of other channels must also be considered.
Search engines are limited for adverse media research
Most businesses and financial institutions conduct due diligence checks by using free internet search engines, such as Google or Bing. Searches are based on names and keywords and, within a few seconds, the search engine returns a list of results from the world wide web that match the search criteria. Analysts, in turn, manually scan the results for potential adverse media mentions. However, these searches typically return a very large number of results that may or may not be relevant and may or may not help the organization mitigate risk.
Boolean search queries using keywords with modifiers such as “and,” “not,” and “or” may return more relevant results, but such searches tend to generate a large number of false positives and may miss significant risks not contained in publicly accessible web content.
When used for adverse media research, search engines suffer from a number of limitations:
- Searches generate a high volume of results with many false positives
- Results often contain multiple hits for the same or similar content
- It can be difficult to determine truly adverse events
- Investigators often have trouble verifying the identity of a subject
- Teams need to overcome language barriers when accessing international content
- Results may miss significant risks not contained in mainstream news sources
In addition, relying on search engines to comply with the requirements of your organization’s CDD policies requires analysts to waste precious time jumping between websites, news media, industry sources, and social media feeds to verify information on a subject.
Replace Boolean searches with natural language processing
Search and filter options based on natural language processing (NLP) can more accurately identify media mentions that matter. Natural language processing gives computers the ability to analyze and understand human language. As a form of artificial intelligence (AI), the technology uses algorithms to help computers understand the way humans speak and write.
NLP technology allows software systems to understand human commands, make sense of content, and cut through the noise to determine what’s important. It’s easier than ever to build efficient content analysis systems that transform your organization’s screening processes.
Conducting adverse media screening involves analyzing and evaluating significant volumes of content. Natural language processing helps risk management teams prioritize and categorize data, tag relevant content, and identify duplicate stories. Eliminating the need to manually evaluate volumes of content saves valuable time and reduces the possibility of introducing human bias or errors.
Automate adverse media screening with artificial intelligence
Natural language processing is just one way that artificial intelligence and machine learning (ML) are optimizing and automating the adverse media screening process. In addition to significantly reducing the time and money spent on identifying, evaluating, and investigating negative news, AI-driven automated screening helps companies maintain regulatory compliance and identify potential financial and strategic risks. It gives risk mitigation and fraud detection teams the tools they need to easily access relevant information and conduct comprehensive, diligent investigations.
Artificial intelligence refers to computers programmed to think like humans, solve problems, and take actions that have the best chance of achieving a specific goal. AI software systems help investigators anticipate challenges and deal with issues as they arise.
Machine learning focuses on developing computer programs that access data and use it to learn for themselves. Over time, ML systems can improve from experience without requiring specific programming.
Machine learning solutions save time by discounting unimportant news, identifying false positives, and recognizing relevant media that needs further evaluation. The technology can determine if a news article actually focuses on the subject of interest or just mentions the subject in another context. It also speeds the review process by highlighting keywords and important terms.
Of course, even the most advanced AI technology requires some manual processing. Risk and compliance teams set and update parameters in the system, configure the tools, customize filters, monitor information, and review the results.
Quickly identify and evaluate high-quality, relevant content
Artificial intelligence and machine learning allow organizations to develop automated adverse media screening processes that deliver accurate, detailed risk management profiles of customers, prospects, and business partners.
With these tools, you can quickly prioritize relevant news articles and:
- Remove multiple hits for the same content. Media searches are large in volume, often with similar or the same stories showing up in the results. AI technology removes duplicates and similar stories from the start of your investigation, giving you relevant, accurate, and transparent search results that are easy to evaluate.
- Discern truly adverse events in search results. AI media screening technology can also help your team discern truly adverse events in search results by providing a relevancy rating that reflects the confidence level of the match and links to the underlying source — so you can verify the identity of a subject and know that “John Smith” is the same John Smith who has just been convicted of money laundering.
Many businesses lack clear procedures for handling adverse media incidents that involve criminal allegations without convictions. The absence of coverage in leading national or international news sources may not be enough to discount an allegation. Being accused of a financial crime is reason enough to conduct further CDD because a client’s bad reputation can pose risks to your organization.
Companies must determine whether the allegation, if proven true, could pose a risk to the organization, or if the subject may be involved in money laundering, terrorist financing, or other criminal activity. Further CDD is warranted if screening identifies a plausible link between the alleged conduct and the subject of the investigation.
- Filter and categorize content to focus on what your team really cares about. Not all adverse media is the same. AI technology with customizable settings automatically filters and tags risks into different categories according to a subject’s profile, allowing analysts to assess the true level of risk. Analysts can also tailor searches and add or delete criteria based on risk considerations, including corruption, financial fraud, money laundering, terrorist financing, cybercrime, drug trafficking, sanctions exposure, bribery, regulatory misconduct, or organized crime.
Access international content and overcome language barriers
A comprehensive program to screen customers against negative news requires full coverage of the international media landscape.
Some countries and regions around the world experience higher levels of money laundering, terrorist financing, financial crimes, and other criminal activity. While many adverse media screening tools focus on the world’s leading media sources and languages, coverage of specific languages in high-risk regions is limited. Moreover, international media often report incidents in local languages, making it difficult for English-speaking investigators to identify and screen for negative news.
Effective screening also takes into account less significant events that may only generate a handful of postings by local or city media. While international news agencies may not pick up these stories, negative news at all levels can pose risks to your organization. Events that occur in remote areas may not reach larger news agencies until days after the initial local media reports. Some events may never reach larger mainstream news agencies.
Advancements in natural language processing and machine translation allow English-speaking analysts to monitor international media at all levels and investigate adverse media mentions in foreign languages. Tools also support automated alias checks in countries where a subject’s name in English may have multiple aliases.
Search beyond the news media
While mainstream news coverage focuses on significant stories of interest to the public, the news media may not report every risk-related piece of information about your customers. Screening press releases, reports by non-governmental organizations, third-party sources of structured data, and notices published by law enforcement agencies, tax authorities, and government agencies can often reveal hard-to-find risk information.
Including open source public records in your searches ensures you are accessing detailed information from a wide range of reliable sources, including the Federal Communications Commission (FCC), Office of Foreign Assets Control (OFAC), drug crimes, trafficking, white collar fraud, sanctions watchlists and blacklists, and more.
Establish an ongoing adverse media review program
Ongoing monitoring is vital for a successful screening program. In addition to ensuring that information is regularly updated and remains accurate, it helps organizations recognize patterns of potentially risky behavior and develop a clearer, more accurate understanding of customers over time.
With intelligent machine-based screening solutions, you can schedule batch screenings of customers as often as needed. For example, you may decide to screen high-risk customers every week, while medium- and low-risk customers require less frequent monitoring.
Assigning a risk score to a customer during the onboarding process won’t necessarily determine how that risk might change in the future. Continuous screening allows analysts to track breaking news and developing stories around the world, quickly identify risk events, and upgrade or downgrade risk scores — all in real time.
Proactively reviewing adverse media gives companies the information they need to respond quickly to sudden changes in activity. In some cases, ongoing monitoring can even prevent a crime before it happens.
Automate risk management and regulatory compliance processes
For organizations that rely on a search engine to conduct adverse media checks — and for those that depend heavily on manual processing — moving to an automated solution can greatly enhance and streamline your due diligence program. Technology powered by artificial intelligence, machine learning, and natural language processing can help you maintain compliance with national and international KYC, AML, and CFT laws.
Augmenting your compliance platforms with innovative digital tools is a key step in taking a risk-based approach to customer relationships and protecting the value of your brand.
Regulatory agencies expect businesses to perform comprehensive CDD checks as part of a risk management process. AI screening technology is the solution to expanding and simplifying your organization’s risk management program and workflow. Advanced automated solutions allow you to easily filter through extensive amounts of data, identify who should be screened and how often, and review and evaluate results to identify any potential risks.