AI, Deepfakes, and the New Frontier of Fraud: What UK Businesses Need to Know in 2025

Once imagined only in sci-fi, deepfakes are now a very real threat to UK businesses. The rise of generative AI has reshaped industries across the board – including, unfortunately, the business of fraud. Today’s scammers no longer rely on poor grammar and obvious phishing emails. Instead, they create powerful AI tools to mimic real voices, fabricate life-like videos, and generate fake documents and identities that are almost indistinguishable from the real thing.

In 2025, deepfake-related fraud reached an all-time high. The UK’s National Cyber Security Centre (NCSC) recently reported that deepfake content usage in fraud has risen by over 400% in the past 18 months, driven by the accessibility of free or low-cost AI tools. We’ve already seen real-world cases: in one incident, the UK branch of a multinational company lost £20 million after a finance employee was tricked into transferring funds during a video call featuring a deepfaked executive. Believing he was speaking to his real colleague, who was clearly visible on the screen, the employee followed instructions to move the money, completely unaware that the entire video call had been fabricated using AI.

Voice cloning is especially worrying. With just a few seconds of recorded audio, scammers can copy someone’s voice and use it to make convincing phone calls to employees, suppliers, or banks. When combined with compromised email accounts, these scams can slip past even the strongest of internal checks. In today’s environment, double-checking a video call or asking for extra verification isn’t seen as being overly cautious; it’s seen an an essential step in protecting company assets and preventing costly fraud.

Which Types of Merchants Are Most at Risk?

The landscape of fraud is rapidly evolving, and no merchant is immune; however, the tactics used by fraudsters vary depending on the industry. For e-commerce retailers, especially those offering high-value goods, fraudsters are using synthetic identities. This is a combination of real and fake data, to open accounts and place orders that later become disputed or charged back. Subscription services are facing growing threats from bot-generated sign-ups and fake account activity, often used in conjunction with AI tools designed to bypass fraud filters.

Digital goods and service businesses are especially vulnerable due to the irreversible nature of their products. Once a download is complete or a service is given, reclaiming lost revenue is extremely difficult if the transaction turns out to be fraudulent. Fraudulent returns, a growing trend in retail, are also being supercharged by AI, with fake receipts, doctored ID scans, and even convincing chatbot impersonations used to pose as customer service staff.

Even traditional brick-and-mortar businesses aren’t completely safe. Many now have online components, such as click-and-collect or delivery services, that are being exploited through account takeovers or fake refund requests. The broader use of QR codes, digital wallets, and contactless payments opens up new avenues for sophisticated attacks.

What AI-Driven Tactics Are Targeting Merchants Right Now?

AI is not just creating new scams – it’s also improving the old ones. Phishing emails, for example, used to be easy to spot due to clumsy language. Now, language models can generate grammatically perfect, personalised messages that convincingly mimic a company’s tone and formatting. Some even target specific employees based on LinkedIn profiles and past correspondence.

Fake reviews are another growing concern, especially for online merchants. AI can produce hundreds of reviews that appear authentic to both customers and algorithms. This not only damages a company’s credibility but could soon place merchants in breach of the UK’s Digital Markets, Competition and Consumers Act, which took full effect in April this year, and aims to clamp down on misleading online practices.

One of the most dangerous tactics is synthetic identity fraud. By blending AI-generated imagery with real-world credentials (for example, stolen National Insurance numbers), criminals create entirely new identities that can pass through Know Your Customer (KYC) processes. These identities are then used to open fraudulent accounts, build fake credit histories, and conduct scams over an extended period of time, before completely vanishing without a trace.

When these scams result in unauthorised transactions or disputed purchases, the burden often falls back onto the merchant in the form of chargebacks. For many, the transaction seems legitimate at the time, but by the time the fraud is uncovered, the product is long gone, and the liability has shifted onto them.

How Are AI-Powered Scams Changing the Nature of Chargebacks?

In the past, merchants could often spot chargeback fraud through looking at patterns. For example, repeat customers, common geographies, and suspicious order behaviour. But with AI, fraudsters are becoming better at mimicking real consumer behaviour, making it much harder to detect fraud before fulfilment. This poses a serious challenge to fraud teams and dispute resolution departments.

AI-generated scams blur the line between first-party and third-party fraud. Was a purchase made by a legitimate customer who later claimed fraud (known as friendly fraud), or was their account hacked, and compromised using AI-enhanced tactics? Proving honesty is increasingly difficult, and card issuers often err on the side of the cardholder, again, leaving the merchant to foot the bill.

This puts pressure on internal chargeback teams to investigate more deeply and document evidence more thoroughly. As cases become more complex and volumes increase, many teams are stretched thin. Add to that the short timeframes for responding to disputes, and merchants find themselves fighting an uphill battle.

While not a silver bullet, partnering with a chargeback solution provider can offer much relief in this situation. Such companies bring advanced tooling, AI-powered analytics, and hands-on expertise to help merchants dispute chargebacks more effectively. Some even use machine learning to identify early fraud indicators, giving merchants a much better chance of spotting scams before they escalate into revenue loss.

What Can Merchants Do to Stay Ahead of the Curve?

Staying ahead in 2025’s fraud environment means taking a layered, proactive approach.

Stronger Verification Checks – Merchants should strengthen their identity verification protocols, especially during onboarding or high-value transactions. This could include things like biometric checks or document verification that goes beyond basic ID scans.

Reinforce Account Security – Multi-factor authentication (MFA) is another vital tool, especially for employee accounts and customer logins. While not foolproof, MFA adds a crucial barrier that AI-using attackers often struggle to get past, especially when used in conjunction with behavioural biometrics, which analyse how users interact with a site or app in real-time.

Staff Training – Staff training must always continue to evolve. Fraudsters are now impersonating CEOs, suppliers, and even colleagues with unnerving accuracy. Regular internal briefings, role-based scenario testing, and clear instructions on what to do, can help staff identify and respond to unusual requests before any damage is done.

Collaborate with Other Merchants in the Industry – Merchants should partake in industry collaboration. Sharing threat intelligence, especially around emerging AI scams, can help create a collective defence against fraud. There are many are valuable spaces for this kind of cooperation.

Specialist Chargeback Support – When it comes to managing chargebacks, it may be time to assess whether your internal team has the proper support. As AI-driven scams continue to complicate the fraud landscape, partnering with a dedicated chargeback specialist can alleviate the strain within operations, boost recovery rates, and free up resources to focus on growth and innovation instead of disputes.

Where Do We Go From Here?
The fraud landscape in 2025 is smarter, faster, and far more convincing than it ever has been in the past. As AI tools become more powerful and accessible, the line between authentic and fake blurs, making life significantly harder for merchants. But with the right strategies in place, from improved identity checks to smarter dispute handling, businesses can still stay one step ahead.

Whilst no single measure offers complete protection, a well-rounded approach that blends prevention with effective resolution is essential in today’s evolving landscape of fraud. This means not only investing in tools and training to stop fraud before it happens, but also having a robust process in place to manage the fallout when it does. For merchants processing high volumes of transactions, where even a small percentage of fraud can make a big impact, partnering with a chargeback solutions provider can offer a quiet but critical advantage. With the proper support, businesses can reduce operational strain, improve dispute outcomes, focus their efforts on growth and face the future with confidence.