Deepfake of NatWest CEO: New AI Scam Shakes Banking

What happened?

On July 3, an AI-generated image showed NatWest CEO Paul Thwaite alongside journalist Emily Maitlis in a fake BBC studio. The image, widely shared on X, included the text: "Live, Emily Maitlis raised NatWest CEO's salary, sparking strong backlash." NatWest confirmed to CityAM that the image is a deepfake and is working with platforms to remove it. A spokesperson stated: "They are another reminder of criminal activity designed to defraud people, especially the vulnerable." The image references the 33% salary increase Thwaite received since 2024, reaching a £6.6 million package last year—a real fact that scammers exploited to lend credibility to the deception.

Why is this important?

This case is not isolated. Bank of England Governor Andrew Bailey had already been a victim of a violent deepfake alongside Nigel Farage. According to Vyntra's 2026 report, deepfake has become a $400 billion annual industry, driven by increased computing power and generative AI. Scammers can create fake videos, images, and voices in under five minutes, compared to 16 hours previously. Deepfake technology has evolved rapidly: in 2023, deepfakes accounted for only 1% of financial fraud; by 2026, they exceed 12%, according to cybersecurity firm Sensity. The combination of open-source tools like Stable Diffusion and voice models like ElevenLabs allows anyone to generate convincing fake content without advanced technical knowledge.

The NatWest case adds to a series of similar attacks on banking figures. In February 2026, a deepfake of HSBC's CEO circulated on WhatsApp offering fake loans. In April, a fake video of Santander's chairman appeared on YouTube promoting a cryptocurrency investment scam. The frequency of these attacks has doubled every quarter since 2025, according to a report by the British Bankers' Association. Banks have invested over £500 million in deepfake detection systems, but scammers adapt quickly: high-quality deepfakes now include natural blinking and synchronized lip movements, making them difficult to identify even for automated tools.

Consequences for banking and users

Deepfakes not only damage executives' reputations but are used for investment scams, identity theft, and employee fraud. In the NatWest case, the fake image could have induced people to invest in fraudulent schemes or share banking details. Banks are strengthening their verification systems, but the sophistication of deceptions outpaces many defenses. For example, in March 2026, a Barclays branch employee transferred £1.2 million after receiving a deepfake voice call imitating the CFO. The voice was generated from just 30 seconds of audio extracted from a public meeting on YouTube.

For the average user, the recommendation is to distrust any content that seems too good or alarming, always verify through official sources, and report any suspicion. However, digital education remains insufficient: an Ofcom 2026 survey revealed that 45% of Britons cannot identify a deepfake, and 30% have unknowingly shared a fake video. Social platforms like X, Facebook, and YouTube have implemented AI-generated content labels, but their effectiveness is limited. In the Thwaite image case, the original X post accumulated over 50,000 views before being removed, and it still circulates in private Telegram groups.

What readers should know

Deepfake technology advances faster than regulation. Social platforms struggle to detect and remove such content before it goes viral. Meanwhile, scammers refine their techniques. Digital education and critical thinking are the best defenses. If a message, call, or video asks for money or personal data, verify through another channel before acting. Authorities recommend setting verification passwords with family and colleagues, and using multi-factor authentication apps that do not rely on voice or video.

Regulation is catching up: the EU's Digital Services Act already requires platforms to label deepfakes, and the UK passed the Online Fraud Act in 2025, imposing fines of up to 10% of global revenue on platforms that fail to remove fraudulent content within 24 hours. However, enforcement is uneven. Meanwhile, the Bank of England has created a working group with major banks to share intelligence on deepfake threats. On an individual level, awareness and caution remain the first line of defense. As the NatWest spokesperson concluded: "The best protection is doubt. If something seems suspicious, it probably is."