By Donald McElligott, VP, Compliance Supervision, Global Relay
Over the past decade we have seen a surge in the capabilities of AI, with the technology outpacing and outperforming itself. However, despite the rapid pace of innovation, highly-regulated financial services organizations are still extremely hesitant to adopt AI-powered strategies across their operations – specifically in communications surveillance. What we have observed is that even more surprising than the number of firms who are not taking advantage of this technology is the number of firms who have no plans of ever integrating AI into their everyday workflows – further highlighting the gap in AI-powered surveillance despite its ability to expedite surveillance processes and spot threats before they wreak havoc.
Last year alone boasted a record number of recordkeeping and off-channel communications enforcements against large financial institutions, which included hefty price tags. For many firms, this could have been mitigated with proper AI surveillance protocols in place. To make matters worse, these fines do not seem to be slowing down anytime soon, as the SEC doled out more than $63 million in combined fines to 12 financial firms for recordkeeping failures in January of this year. These ongoing penalties should play as a wake-up call to financial firms to close the gap on AI adoption, as we are on the cusp of a time when genAI will completely change how communications surveillance works. This will mark a significant industry shift, especially given that communication monitoring has remained fundamentally the same for the last 10 years. One way to close this gap in AI adoption is to explore and learn from how surveillance teams can and have been integrating the technology into their businesses.
Implementation of AI-powered audio recordkeeping is on the rise
While it is evident that many financial firms are reticent to go all-in on AI adoption, a lot of firms have made implementations, at least to some degree, to benefit from their risk identification capabilities, ability to reduce false positives and demonstrated proficiency in transcribing voice communications.
These advancements come at a good time, as voice regulation enforcements have been sweeping the US at an alarming rate. In September 2024, the CFTC fined a firm $650,000 for ‘recordkeeping deficiencies and failure to obtain customer authorizations before entering trades for customers.’ This lack of audio violated the CFTC Regulation 1.31 and 1.35 related to recordkeeping obligations to capture oral communications. This fine reinforces the importance regulators place on voice capture as part of communications compliance. In order to prevent future fines and ensure complete retention of all audio and voice communications, especially ones made through third-party platforms such as Zoom, Skype and Microsoft Teams, financial firms should outline and adhere to rigorous procedures internally across the compliance workflow, which can be accomplished with the help of AI.
Firms have also seen resource-saving benefits in using AI for the efficient multi-dimensional risk identification without generating false positives. GenAI, working in tandem with traditional AI models, offers compliance teams the revolutionary ability to reduce noise from irrelevant keywords and, more importantly, to identify true risks by understanding the context of messages.
Leveraging a solution to capture all communications accurately and the ability to store this data for regulators as needed will prove a game changer to firms, especially as the technology continues to evolve and advance.
Explainability outweighing accuracy?
While the pros do seem to outweigh the cons in regards to AI adoption, firms are still hesitant to integrate the technology into their operations. This could be due to a lack of understanding around how AI can be successfully used as a surveillance tool, a lack of budget to implement the technology and data security concerns.
Explainability, the ability of AI systems to provide clear and understandable explanations for their actions and decisions, making the information understandable to users, has been a barrier for AI adoption within surveillance and compliance. However, with the latest AI advances offering significantly better results and far outperforming traditional approaches, the prioritization of explainability over accuracy may not be the ideal approach.
Proceeding towards AI integrations with caution
Additionally, transparency comes into the mix with firms yearning to feel confident that the models employed within their surveillance systems are carefully assessed and authenticated. Understandably, firms want to see how AI-enabled systems are managed, how they’re trained, and how information is documented. It is essential for firms to understand how they will operate so they feel a trustworthiness that the AI technologies they have invested in are being implemented correctly.
Ultimately, organizations need to know how their data is processed, where it is stored and who has ultimate ownership over that data. Given the multitude of uncertainties and ethical considerations surrounding AI, it is understandable that firms remain cautious on the matter.
Regulators and financial examiners have taken the stance that firms adopting AI solutions need to understand the technology they employ. It is important they know how their AI solution works to reassure the financial examiners they have the skills to properly use the technology. That adds some responsibility to the AI vendors. Firms are not only buying the AI solution; they also should make sure they are receiving the proper training and information to fully understand the technology being deployed.
What’s ahead for AI in business communications surveillance?
Regulators have remarked that it’s undeniable that AI is and will continue to have a huge impact on industry operations. AI has proven a crucial tool used for various surveillance systems, specifically communications. Since the introduction of genAI, the technology has demonstrated the ability to radically transform and streamline communications surveillance processes by decreasing false positives and identifying risks within written communications. Due to these strong use cases and benefits, at the start of 2025 the current administration released an executive order that would revoke certain existing AI policies and directives that act as barriers to American AI innovation, clearing a path for the United States to act decisively to retain global leadership in artificial intelligence.
As both risks and benefits are taken into account, one thing is for certain. AI is here to stay, so it is essential for financial firms to prepare for the eventual mass adoption of this technology and to try to close the gap on effectively implementing it into their workflows.
