AI and fraud: opportunities and risks in the digital age

AI knowledge development grows; 'Focus on Fraud' webinar highlights prevention of AI-related fraud and opportunities of AI for organizations.

The evolution of AI offers enormous opportunities but also brings risks. Due to rapid developments and still limited knowledge, there is a risk of fraud. Two crucial questions are central here: does AI make committing fraud easier, and can we use AI to detect fraud? Financial professionals, CEOs, and board members are immersing themselves in the rapid developments of AI, given its great importance for their daily responsibilities.

Diana Matroos, journalist and presenter of 'De Big Five' on BNR Nieuwsradio, served as the hostess in the webinar alongside Auke de Bos, an executive at EY accountants. Together, they engaged in conversation with experts to increase awareness of AI.

AI: opportunities and fraud risks

Igor Mikhalev: “In a world where technology advances at an unprecedented pace, AI sometimes worries me. While we aim for creativity and innovation, AI inevitably creates opportunities for people with illegal intentions. The results of our survey show high concern over misinformation and criminal activities, but also recognize efficiency and cognitive improvements as opportunities. I wonder if we can ever fully solve these problems, or if it's a perpetual balance that we must strive to maintain. With advancements in GenAI, we see an increase in sophisticated fraud, from hyper-personalized text scams to voice attacks and deepfakes. Hackers are now capable of carrying out attacks on a large scale.”

Despite 52% of consumers in the survey thinking they can detect a deepfake video, there's a strong chance that this confidence is misplaced. Igor: “The impact of such technology on society and business is profound, and as the proverbial flywheel of innovation spins faster, companies must become more agile in identifying and understanding these threats. They must also proactively deploy technology against such attacks, as human ingenuity alone is insufficient to prevent the advanced level of attacks we now face. Companies need to be agile and smart in detecting and understanding these threats, and proactively use technology to counter them. We need technology that matches the scale of these attacks.”

Critical assessment

It seems to be a complex issue, especially as we see an increase in sophisticated fraud. Igor: “With the current climate of misinformation, this can also lead to trust issues. As an individual, I am very interested in this problem. The enormous amount of data and information being generated, blurs the line between truth and falsehood. We see multiple versions of the truth, some based on what people have actually said, some generated by AI, and all subject to emotionally charged interpretations. This leads to different versions of the truth, depending on who you ask. As business leaders, we need to educate our stakeholders on how to critically assess information. We also need to redefine our understanding of truth in the context of this information overload.

As AI continues to evolve and associated fraud becomes more sophisticated, we must remain vigilant and adaptive. It is essential to invest in education, advanced detection technologies, and robust policy measures to mitigate these risks. We are in a constant race to outsmart cybercriminals, and while our detection methods are advancing, they do not keep pace with the rapid progress of AI-powered threats. As we navigate the unknown waters of AI, we must be willing to face the challenges head-on while embracing the opportunities it offers. The future of AI is a canvas of unlimited possibilities, but it requires a careful and sustainable approach to ensure it benefits all of humanity.”

Human factor

To understand the impact of AI on individuals, we must protect people and focus on leveraging AI for opportunities and anticipating fraud risks. It is important that we prepare people and stimulate critical thinking. Bernadette Wesdorp: “The research indicates that 35% of organizations have not yet deployed AI, while others are already using it for, for example, chatbots. You would expect that the percentage not using AI would have decreased by now, as it is sometimes unknown that AI is already being used in some applications. With the rapid adoption of GenAI, this percentage will likely quickly drop to zero. You can no longer avoid using AI.”

Navigating AI legislation

With the advent of the AI Act from Europe, it is time for organizations to understand the impact of this legislation and to take stock of where AI is used in an organization. Bernadette: “For Europe, the text has been published and by the end of the year, the law will apply to systems that are truly prohibited. This means that organizations now need to think about their approach and the impact of the legislation.”

Understanding AI activities

Identity and access management, in other words who has access to which systems, is not a new issue, but AI brings both new and existing risks to light. Organizations must be aware of their current measures and what may be additionally needed. Simple solutions such as two-factor authentication can prevent many problems and fraud, despite the desire for ease of use and quick access. Bernadette: “AI is a multidisciplinary theme that affects all facets of an organization, from data and IT to legislation, compliance, and legal aspects. It is crucial to tackle AI-related fraud from different angles, for example through a multidisciplinary AI board or ethics board that evaluates AI initiatives. Organizations must understand how they deploy AI and continuously monitor it, given the rapid developments and the different roles within AI legislation, such as developer or user. Insight into one's own AI activities is incredibly important.

With the arrival of the AI Act, it is time to take action and assess the impact on your organization. Often there is already a good governance structure in place. It is important to identify where the shortcomings are and what still needs to be done. Take stock of what your organization is already doing and build on this to integrate both new risks and opportunities. You don't have to start from scratch, but it is essential to investigate which steps you need to take now. It is crucial that the responsibility for AI is taken from the top of the organization, such as the board of directors. The most important thing is how you respond to unforeseen events; full preparation is not possible. Focus on the future, learn from experiences, and ensure a quick response.

It is essential to refine our 'AI antenna,' which means we must critically assess whether we can trust the information we receive. As directors, who constantly make decisions, we must be vigilant for the reliability of the data presented: is it true or not? This critical attitude is not only important for ourselves but also something to share and emphasize within our organizations. The human element is indispensable alongside AI, which can both combat fraud and add value. It is an era of opportunities, but also of vigilance and responsibility.”