Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

Practical AI use cases even skeptics should embrace

With proper implementation, AI tools can be a valuable addition to any company’s security and anti-fraud stacks.

Practical AI

[Adobe Stock – HockleyM1/peopleimages.com]

Profile photo of Rafael Lourenco

EVP & Partner · ClearSale

Despite the rapid growth of AI capabilities—and the eagerness of many companies to experiment with AI—not every business leader is ready to work with AI.

Questions remain about AI’s reliability and cost-effectiveness. But I believe every business that cares about data protection, customer relationships, compliance, and reputation should use AI in their security and fraud prevention technology stacks to counter the growing wave of AI-powered attacks.

SYNTHETIC IDENTITY FRAUD, SUPERCHARGED

Identity fraud is used to link to specific real people, but synthetic identity fraud blends personally identifiable information (PII) from multiple people to create new identities, using AI to fill in the information gaps and fabricate realistic-looking verification documents. As the “fastest-growing financial crime in the United States,” per Deloitte, synthetic identity fraud has become a time bomb in financial institutions’ customer databases.

Auto loan providers are frequent targets for this type of fraud, losing $7.9 billion in 2023. While lenders and banks are seeing the most synthetic identity fraud so far, 46% of all businesses were hit by it in 2022, including government agencies. Standard online document review systems and expert analysts can be deceived by the quality of these synthesized identities. A more effective approach uses AI to check online loan and new account applications against real-time data sources and behavioral biometric markers that indicate identity theft and bot-assisted fraud attempts.

SPEAR PHISHING AT SCALE

Stolen data fuels synthetic identity fraud, and it’s valuable to criminals in other ways. That’s why they devote so much effort to researching valuable targets in companies and crafting messages to trick those people into sharing login credentials, paying a fake invoice, or diverting an employee’s payroll.

Now, with AI, scammers can quickly gather data from company websites, sales databases, and platforms like LinkedIn, creating phishing messages that are “more advanced, harder to spot, and significantly more dangerous” than before. AI also broadens access to spear phishing. For example, security researchers recently found they could use a free GenAI tool and a stolen email address to generate and send phishing messages en masse.

As with synthetic identity fraud, analysis using real-time data and behavioral biometric patterns to stop these attacks is only practical with AI tools.

CONVINCING BRAND IMPERSONATIONS

Combining AI image generation with spear phishing strategies lets criminals launch impersonation attacks aimed at consumers, who believe they’re interacting with a trusted brand or government agency. In reality, they’re giving personal information, payment data, or login credentials to fraudsters. Consumers lost more than $1 billion to brand imposters in 2023.

AI-based tools can help companies continuously monitor the web and social media for imposter ads, accounts, and websites. AI can also automate the documentation, reporting, and takedown of imposter accounts, which can be an ongoing battle (full disclosure: ClearSale offers these solutions).

ASSESSING YOUR COMPANY’S AI READINESS

The age of accessible AI is another round in the ongoing cat-and-mouse game between businesses and criminals. Now that everyone can access powerful AI tools, companies without similar defenses will be increasingly vulnerable to attacks.

To get real value from AI security tools, companies need to meet basic readiness requirements. Start with a policy that defines how the initial AI use case. A bank might start with new account application screening, an e-commerce brand might start with order screening, and a manufacturer might focus on preventing business email compromise. Determine which employees will work with these tools, how they will be trained, and how the AI solution’s performance will be monitored. The AI model will also need unified, standardized data to ingest and analyze.

Defining and implementing AI guardrails is also critical. For example, McKinsey’s AI guardrail taxonomy recommends controls to govern AI results’ appropriateness, accuracy, and alignment with user expectations, as well as regulatory compliance. The most important guardrail is validation, where results are checked to ensure they contain required elements and don’t contain any off-limits information. The validation step should be backed by an expert human review of any results that raise flags.

WHAT TO LOOK FOR IN AI TOOLS

Evaluating AI security options can be a challenge not only because the technology is evolving fast, but also because a lot of vendors are jumping on the AI bandwagon with varying results. Look for AI-native solutions built for a specific use case and industry, as they are likely to be more effective and easier to use than legacy solutions with “bolted-on” AI components. Quality solution providers will be able to share data on the real-world effectiveness of their offerings and discuss their compliance and data security practices.

The humans behind the tools matter, too. AI solution providers that respond to customer questions and offer suggestions, and those that are continuously anticipating new developments in AI, are more likely to help their customers stay ahead of crime and fraud trends as AI capabilities expand.

FIGHTING AI WITH AI

AI tools can be a valuable addition to any company’s security and anti-fraud stacks. On its own, the potential volume and speed of AI-enhanced attacks would be enough to recommend AI solutions to quickly assess and respond to attacks and fraud attempts.

The fact that AI allows these attacks to be ever more precisely targeted and persuasive is another reason to bring on AI solutions that can identify anomalies humans might miss—even in a business that isn’t using AI anywhere else.