AI-Powered Disinformation Poses Escalating Threat of Bank Runs, Study Warns
The rapid advancement of artificial intelligence (AI) has ushered in a new era of sophisticated disinformation campaigns, posing an escalating threat to the stability of financial institutions, according to a groundbreaking study conducted by UK research firm Say No to Disinfo and communications firm Fenimore Harper. The study reveals how AI-generated fake news and manipulative social media content can trigger bank runs, urging financial institutions and regulators to bolster their monitoring efforts and proactively mitigate these emerging risks.
The study highlights the alarming ease with which malicious actors can leverage generative AI to fabricate convincing fake news articles and memes, disseminating them rapidly through social media platforms using paid advertising. These fabricated narratives can sow seeds of doubt and panic among bank customers, creating a domino effect that can lead to widespread withdrawals and ultimately, bank failures. The researchers demonstrated the potential impact by exposing UK bank customers to AI-generated content designed to undermine confidence in their financial institutions. The results were striking, with one-third of participants indicating they were "extremely likely" to withdraw their funds after viewing the fabricated content, and an additional 27% expressing a "somewhat likely" inclination to do the same.
The speed and scale at which disinformation can spread through social media, coupled with the convenience of online and mobile banking, create a volatile environment where mass withdrawals can occur within minutes. The study estimates that a mere £10 spent on social media advertising to amplify fake content could trigger the movement of up to £1 million in customer deposits. This estimate is based on average deposit amounts, advertising costs, and projected reach of social media campaigns.
The 2023 collapse of Silicon Valley Bank, where $42 billion was withdrawn in a single day, serves as a stark reminder of the vulnerability of financial institutions to social media-fueled panic. The G20’s Financial Stability Board has also sounded the alarm, warning that generative AI empowers malicious actors to orchestrate disinformation campaigns that could precipitate financial crises, including flash crashes and bank runs. The Say No to Disinfo study emphasizes the urgency of addressing these risks, urging financial institutions to integrate media and social media monitoring with withdrawal tracking systems to detect and respond swiftly to emerging threats.
In response to the study’s findings, Woody Malouf, head of financial crime at Revolut, a London-based fintech company, affirmed the importance of real-time monitoring for emerging threats, both within their customer base and across the broader financial ecosystem. Malouf stressed the need for preparedness, emphasizing the potential for such industry events, though unlikely, to occur. He also called upon social media platforms to assume greater responsibility in curbing the spread of malicious information. Other major financial institutions contacted for comment, including NatWest and Barclays, either declined or did not respond.
While regulators express growing concern about the overall implications of AI for financial stability, banks generally maintain an optimistic outlook on the technology’s potential benefits. UK Finance, an industry body, acknowledges the ongoing efforts by banks to manage and mitigate AI-related risks, while emphasizing the role of regulatory authorities in addressing the broader financial stability challenges posed by AI. The study’s release coincided with an AI Summit in France, where the focus shifted significantly from risk management, which dominated previous summits, to promoting the wider adoption of AI. This evolving landscape underscores the critical need for a balanced approach that harnesses the transformative potential of AI while proactively addressing the associated risks to safeguard financial stability.