AI Misinformation Could Lead To Mass Bank Withdrawals

A new British study warns that artificial intelligence-generated disinformation circulating on social media is increasing the risk of bank runs, urging financial institutions to enhance their monitoring efforts, as first reported by Reuters.

The research, published by UK-based Say No to Disinfo and communications firm Fenimore Harper, highlights how AI makes disinformation cheaper, faster, and more effective, a wider range of actors—including those driven by financial, ideological, or political motives—could exploit this vulnerability.

The ease of online banking and rapid money transfers further increase banks’ exposure to such risks.

To analyze the potential impact, the researchers created an AI-generated fake news campaign targeting banks’ financial stability. False headlines were designed to exploit existing fears, using doppelganger websites and AI-generated social media content.

The researchers simulated large-scale amplification, generating 1,000 tweets per minute at minimal cost. A poll of 500 UK customers showed that after exposure to the disinformation, 33.6% were extremely likely and 27.2% somewhat likely to move their money, with 60% inclined to share the content.

Based on average UK bank account balances, a single disinformation campaign could move £10 million, with the cost of shifting £150 million as low as $90–$150.

Despite the speed and ease of such attacks, the researchers say that financial institutions remain unprepared. Banks lack disinformation specialists, proactive monitoring, and crisis response plans. Current security measures focus on cyber threats while neglecting AI-driven influence operations.

AI-enhanced disinformation has the potential to destabilize the financial sector by eroding trust and triggering large-scale withdrawals. Without proactive measures, the researchers claim these campaigns could cause widespread economic damage.

Reuters notes that concerns over AI-driven disinformation follow the 2023 collapse of Silicon Valley Bank, where depositors withdrew $42 billion in a single day.

Regulators, including the G20’s Financial Stability Board, have since cautioned that generative AI could exacerbate financial instability, warning in November that it “could enable malicious actors to generate and spread disinformation that causes acute crises,” such as flash crashes and bank runs, as reported by Reuters.

While some banks declined to comment to Reuters, UK Finance stated that financial institutions are actively managing AI-related risks.

The study’s release coincides with an AI Summit in France, where leaders are shifting focus from AI risks to promoting its adoption.