Adverse Media Screening 101 – An introduction to Adverse Media
What is Adverse Media?
Adverse media can be described simply as news published in the media that has negative connotations. In the commercial world, it is news related to a particular entity, often an individual, a company, or a brand. Adverse media can be found in a wide variety of places— online outlets, traditional media such as newspapers, magazines, TV, radio, and to an extent social media platforms.
Why do companies need Adverse Media Screening?
Regulations such as the Customer Due Diligence Requirements for Financial Institutions set out by The Financial Crimes Enforcement Network (FinCEN) and The Risk Factors Guidelines under the EU’s 4th AML directive have meant that financial institutions and other corporate entities have a legal requirement to carry out significant Customer Due Diligence (CDD) checks on anyone they do business with as part of the following risk processes.
- Know Your Customer (KYC)
- Financial Crimes
- Anti-Money Laundering (AML)
This focus has meant that, along with monitoring and cross-checking watchlists and blacklists, organizations and financial institutions are also leveraging international media content as an additional data source for screening and mining activities.
Effectively Tracking Negative News
It’s thought that upwards of exabytes of data/content are published online every day. This makes finding what matters in all of that unstructured data next to impossible, for two reasons:
1. The sheer volume of content
2. The complexity of this content
Machine Learning (ML) and Natural Language Processing (NLP) advancements have made it easier than ever before to build intelligent and efficient content analysis pipelines that can transform an organization’s adverse media search and tracking process.
Automation and Augmentation
Typically traditional adverse media screening is based on inefficient manual processes and in some cases a rudimentary Google search.
However, by leveraging AI-powered solutions time consuming and analyst heavy tasks based on discovery and investigation can be heavily augmented and even fully automated to significantly reduce the time spent on identifying and interrogating adverse media mentions.
Proactive and On-going Search
In CDD and KYC use cases ad-hoc searches are carried out to run checks for negative mentions of entities and affiliates and often a once-off report is drafted to cover the results of the adverse media check at that time. However, assigning a risk score or classification to that entity during an onboarding process won’t uncover how that risk might change over time. For example, the process doesn’t account for on-going monitoring of that particular entity to catch events of interest that may happen throughout the lifecycle of the business or even events that may only be reported after the business relationship has concluded.
Realtime and continuous monitoring of news content using an ML and NLP powered news intelligence platform means analysts can build detailed and flexible entity, sentiment, and category-based searches to uncover both past, present, and future reputational risk events as they unfold. Advanced event detection and topic models can also be effective in identifying adverse media events as they’re reported.
* Keep an eye out for part 2 of this blog where we’ll show you how to build intelligent searches for negative media mentions using our Entity Search, Categorization, and Sentiment Analysis features.
Traditionally, media monitoring and adverse media screening techniques are based on manual or automated searches using complex keyword searches and queries. While these complex queries can be effective as ad-hoc searches the challenges associated with updating and maintaining these searches make them ineffective as part of an ongoing adverse media screening process. Using Boolean-based or even RegEx-based queries can also result in a large number of missed mentions and even more false-positive results.
Search and filter options based on advanced NLP such as Entity Recognition, Categorization, Sentiment Analysis, and Event Detection models mean that content analysis pipelines used in adverse media processes can be far more efficient and accurate in identifying media mentions that matter.
Global and Multilingual Reach
Global coverage of reporting and the media landscape is crucial to a successful monitoring process. While the Western world for the most part, has taken significant steps in the last decade or so in combating Money Laundering, Terrorist Financing, and Financial Crimes, we are yet to see such advancements and commitment from countries in Africa , Asia and the CES region as is evident in the Basel AML Index for 2019. This means that there is a lack of data, monitoring, and reporting of risk events in these regions by the mainstream media. Furthermore, when events are reported they are often in local languages which can make them particularly difficult to identify and investigate. Most monitoring services will have the world’s leading media sources and languages covered, however, coverage of longtail sources and specific languages from significantly risky regions is greatly lacking.
Coverage in multiple languages and a focus on longtail and local sources means that organizations can rely on search and monitoring services to provide a truly global coverage. Using advancements in Machine Translation organizations can also ensure that adverse media mentions in native languages can be uncovered and investigated sufficiently by English speaking analysts.