How Consilient is transforming financial crime prevention with federated AI

Founded in 2020, Washington DC-headquartered Consilient brings together next-gen technology and best-in-class AML and CFT knowledge to power a more secure, dynamic and effective solution for financial institutions.

In the view of Ajit Tharaken, CEO of Consilient, the key challenge that led to the creation of Consilient centers around a broken and archaic AML/CFT system.

He explained, “The global AML and CFT framework is outdated and ineffective in combating modern financial crime. Financial institutions rely on legacy rules-based systems that often lead to challenges such as missing valid alerts and the generation of high false positive rates – 95% in some cases – which overwhelm compliance teams.”

Such legacy rules-based systems can also struggle to adapt to evolving criminal tactics such as smurfing, round-tripping, complex layering schemes, high-risk jurisdictions, rapid movement of funds and behavioral changes. Chiefly, it becomes tough for such systems to keep up with the ever-evolving tech advances that criminals adopt.

Other drawbacks include the costly and inefficient nature of legacy rules-based systems, eating into vast compliance budgets while providing limited detection improvements.

On top of this, Tharaken identified some notable regulatory and operational barriers. “Increasingly stringent regulations are restricting data movement, which is making cross-border intelligence-sharing tough,” he said.

“Regulatory bodies, central banks, and financial institutions also lack a unified approach to tackling financial crime. Add onto this the cost of non-compliance rising, with record AML enforcement actions against banks that are failing to meet regulatory expectations,” added Tharaken.

Such inefficiencies in legacy rules-based systems are enabling criminals to exploit regulatory gaps, claims the Consilient CEO, which makes financial systems vulnerable to money laundering, terrorist financing, illicit trade and fraud.

In addition, there is no collaboration between the public (regulators, central banks, FIU’s) and the private sector (financial institutions). No feedback loop exists, and financial institutions have no feedback on SAR filings.

Consilient’s answer

To meet these challenges, Tharaken believes the solution comes from AI-powered financial crime prevention.

He said, “Consilient was created to fundamentally transform financial crime prevention by deploying innovative Federated Machine Learning. Unlike traditional AML systems that rely on siloed data, Consilient enables cross-institutional intelligence-sharing without moving or exposing sensitive data.

According to Tharaken, Consilient’s federated AI models are able to detect 3x to 5x more previously missed suspicious activity, reduce false positives by up to 80%, learn and evolve with emerging FinCrime trends and operate in full compliance with data privacy laws – ensuring no sensitive customer data is shared or exposed.

Federated machine learning enables financial institutions to collaborate on financial crime detection without ever moving or aggregating their data. Instead, Tharaken explains, AI models are trained locally at each institution and then combined into a federated champion model which continuously improves as it learns from diverse datasets. This also allows for sharing of insights between the public and private sectors.

He also mentioned the technology addresses some of the major regulatory hurdles that prevent financial institutions from sharing intelligence – which makes it a significant change for global AML/CFT compliance.

The evolving regulatory landscape

An evolving regulatory landscape around AML/CFT requirements is also playing a key role in the AI development area. According to Tharaken, governments and regulators globally are tightening AML/CFT requirements, with increased focus on a number of key areas.

Firstly, AI and ML adoption is taking up a considerable portion of this focus. “Regulators are beginning to endorse AI-driven compliance solutions, recognizing their potential to improve financial crime detection,” said Tharaken.

In addition, AI is enabling a shift from reactive to proactive financial crime prevention. In the area of cross-border data privacy laws, with an increasing number of jurisdictions enforcing such laws, compliance solutions must operate without breaching data sovereignty rules.

What are some of the key regulations impacting AI-driven financial crime prevention? Tharaken raised firstly the US AMLA of 2020, which expands AI adoption for transaction monitoring and beneficial ownership reporting.

Another is the 6AMLD in Europe, which increases liability for financial institutions failing to prevent money laundering, and FATF Guidance, which encourages AI adoption but demands explainability in model decision-making.

Completing the regulation list are GDPR – which requires AI models to comply with strict data privacy and accountability rules – and the PIPL of China, a regulation which imposes stringent controls on cross-border data transfers.

Tharaken explained, “As regulators continue modernizing AML frameworks, AI-driven compliance solutions must ensure transparency, auditability, and compliance with explainability requirements.”

A force multiplier

When it comes to the ever-present fight against financial crime, the role AI can play over the long-term is increasing in the minds of the industry’s top brass.

Tharaken highlighted, for example, that AI can enable smarter risk detection. “AI models can detect hidden patterns in financial transactions that traditional systems miss.”

AI also enables dynamic transaction monitoring, reducing reliance on outdated rules-based systems, whilst AI models are also able to continuously learn and adjust to new criminal methodologies, such as synthetic identity fraud or crypto-based laundering.

Furthermore, there is a key benefit for Tharaken that AI-powered compliance can substantially reduce manual review workloads – which saves institutions millions annually.

AI adoption challenges

Despite a fast-paced rise for AI in the financial crime compliance arena, there are still adoption challenges that remain.

An area of particular importance surrounds data privacy and security concerns. “Financial institutions handle sensitive customer data, and regulatory bodies require strict compliance with data protection laws. Traditional AI models require data aggregation, which poses cybersecurity and legal risks,” said Tharaken.

He explained how Consilient’s federated learning models ensure that no customer data is ever moved or shared – which maintains compliance with GDPR, PIPL and other data privacy laws.

There also remains regulatory hesitancy in AI adoption. In the area, whilst many regulators clearly acknowledge the potential of AI, many remain bearish due to concerns surrounds bias and delays.

On bias, AI models must be interpretable and auditable. According to Tharaken, regulatory approval delays also create doubts – with AI adoption often hindered by slow-moving regulatory frameworks.

The solution for these issues, in the mind of Tharaken, is that Consilient itself collaborates with regulators and financial institutions to ensure AI models are transparent, explainable, and compliant with global AML laws.

One of the biggest albatrosses that remains in the AI space is that of institutional resistance. This plays out across various industries – with many organizations bullish on the idea of AI, but still not yet convinced of its role within their own businesses.

Tharaken said, “Financial institutions are risk-averse and hesitant to replace legacy systems. Compliance teams lack AI expertise, making adoption difficult. Our solution to this is that we provide turnkey AI solutions that integrate seamlessly with existing compliance systems without requiring extensive infrastructure changes.”

The future of AI-driven FinCrime prevention

Looking toward the future of AI and its role within financial crime compliance, there are a number of key signposts for firms to take note of. Firstly, there will be more public and private collaboration on AI models.

Tharaken explained, “Regulators, financial institutions, and technology providers must co-develop AI models to create a standardized approach to financial crime prevention.”

There will also be more AI-powered fraud and AML convergence – with AI models integrating fraud detection and AML compliance, creating unified risk management platforms for FIs.

In the midst of the AI explosion, greater regulatory harmonization is going to be inevitable. For Tharaken, governments and regulatory bodies must harmonize AI governance frameworks to enable cross-border AI-driven compliance.

Finally, there will be an increasing rise in AI and blockchain for transparent AML compliance. “The integration of blockchain technology with AI will enhance auditability, traceability, and regulatory oversight of financial transactions,” said Tharaken.

Leading the future

In an age where technology is evolving at a breakneck speed, the need for firms to not only match the moment but also stand out is becoming less of a desire and more of a need. For Tharaken, the financial industry is at a turning point.

He explained, “Traditional AML compliance approaches are no longer sufficient to combat today’s rapidly evolving financial crime threats. AI-powered financial crime detection is not just an innovation, it is a necessity.

“Consilient is leading this transformation by pioneering federated learning AI that enhances financial crime detection, ensures compliance with global regulations, and redefines collaboration in the financial ecosystem. By leveraging AI in a privacy-preserving, regulator-friendly, and highly effective manner, we are building the future of financial crime prevention today,” Tharaken concluded.