AI scams are weakening trust in genuine online offers

AI scams are becoming a problem for genuine offers too

AI-assisted fraud creates a business problem even when a company’s own offer is genuine. As generated scams become more convincing, consumers have less confidence in the signals they use to judge online offers. That uncertainty can make them hesitate or withdraw from digital activity, which means fraud can impose costs on legitimate businesses as well as its direct victims.

National Trading Standards, a government agency, links that uncertainty to genuine businesses being held back when making offers. Its concern points to two connected effects of AI: criminals can produce stronger deception, while legitimate businesses have to reach people who are less certain about what deserves trust. Understanding the business effect therefore starts with how AI changes the production of fraud.

AI is industrialising familiar scams

National Trading Standards describes the change as an amplification of established forms of fraud. It says: “[Our] intelligence shows that AI is not necessarily creating new types of scams; it is supercharging existing ones. Shopping, holiday, investment and romance scams can now be made more convincing, personalised and scalable, enabling criminals to target more people, more quickly and with less technical expertise needed than before.”

Those four categories show where the amplification operates. Shopping, holiday, investment and romance scams already give criminals established ways to approach potential victims, while AI can increase the credibility, personalisation and volume of those approaches. Criminals can consequently combine individually tailored deception with high-volume campaigns.

That higher output comes with a lower expertise requirement. Michael Bichard, chair of National Trading Standards, says AI is “transforming the fraud landscape” because a new generation of organised criminals can create convincing scams with little technical expertise and operate at “huge speed and scale.” Criminal groups can consequently expand their reach without a matching increase in specialist capability.

The lower barrier changes which AI risks deserve immediate attention from business and technology leaders. Existing forms of fraud can become cheaper to strengthen and easier to repeat, so focusing mainly on novel forms of AI crime can miss an immediate source of risk. A familiar scam can become materially more effective when generated content improves its presentation, personalisation and volume.

That greater effectiveness also changes the information available to consumers. A person facing more convincing fraudulent material can place less weight on surface quality when deciding whether an offer deserves confidence. As polished images, video, voices and websites become easier to generate and deploy, production quality becomes a weaker signal of credibility.

These changes reinforce each other throughout the fraud process. Criminals can reach more people faster because the expertise barrier has fallen, while their targets face harder judgments because the resulting material can be more persuasive. The effects therefore extend beyond people who ultimately become fraud victims into the wider environment where genuine businesses make online offers.

The cost of better fakes is broader uncertainty

That wider effect appears in National Trading Standards survey findings on recognition, consumer strain and behaviour. The three measures capture different responses to scams, with each showing a separate part of the trust problem.

Difficulty distinguishing an AI fake from a genuine offer changes the judgment a consumer has to make. Legitimate offers must compete with generated material capable of reproducing more of the signals people previously used to assess credibility. Feeling “overwhelmed” captures a different effect: consumer strain around scam detection rather than the behaviour involved in identifying a fake.

For some consumers, that uncertainty also changes behaviour. Avoiding certain online activity shows that concern can lead to withdrawal, even though the finding does not establish which activities people avoid. Once that happens, the consequences of fraud reach legitimate companies whose prospective customers become more cautious about digital interactions.

For those companies, withdrawal becomes an external cost of fraud because deteriorating confidence can reduce their ability to engage potential customers. Marketing, product, security and fraud teams consequently share an interest in whether the wider online environment gives people enough confidence to act on genuine offers. That shared interest leads to two kinds of response: helping individuals recognise deception and disrupting the systems criminals use to distribute it.

Consumer education is one layer of the response

The first layer focuses on decisions individuals make when fraudulent material reaches them. On 30 September, National Trading Standards launched its AI Scam Watchlist to expose tools and techniques that criminals increasingly use to strengthen scams and to give consumers guidance on protecting themselves. National Trading Standards identified five AI-powered scam techniques that most concern consumers:

Those techniques matter because generated media can make judgments at the point of contact harder. The Watchlist aims to make these methods easier to understand before a consumer encounters them during a possible scam. Bichard acknowledges that people can feel powerless when faced with AI, but argues that prevention remains a critical defence against fraud.

That defence depends on giving consumers information they can apply during an encounter. Bichard’s proposed mechanism is to explain how AI can make deception more convincing and then give people clear steps to protect themselves. He says equipping people with understanding and practical protection will be “an incredibly powerful tool” against AI-enabled scams.

Practical knowledge can improve an individual’s response, while the volume and organisation of AI-assisted fraud also create a system-level task. Criminal infrastructure, accounts and distribution channels determine whether fraudulent approaches reach consumers in the first place. Because AI enables criminals to produce convincing material faster and with less expertise, disrupting those assets becomes the second layer of the response.

Fraud at AI scale also requires disruption at system scale

That second layer operates further upstream. In March, the UK government announced the formation of a £30m Online Crime Centre intended to become a central point for fighting cyber fraud and disrupting the gangs behind what was described as the nation’s most pervasive crime. The Centre is designed to combine information that individual institutions would otherwise see only from their own positions in the fraud ecosystem.

Combining that information requires institutions with different views and powers to work together. Experts from government, police forces, the intelligence community, financial services, mobile networks and large technology firms are expected to work side by side. They will share crime data in real time and build a unified picture of global fraud networks, giving participating organisations a common basis for identifying activity that crosses institutional boundaries.

That shared picture is intended to support direct disruption. The Centre’s remit includes identifying the accounts, websites and phone numbers used by organised cyber crime gangs and taking them down at scale. Consumer education helps when a person encounters a suspicious approach; infrastructure enforcement acts earlier by interfering with the assets and channels used to run campaigns.

The Centre also forms part of a larger government commitment. Its £30m funding was announced as Westminster expanded its strategy to put £250m into anti-fraud activity through 2029. That wider investment places coordinated intervention alongside measures that help individuals identify and respond to scams.

The same need for coordination appears in John Herriman’s view of consumer protection. Herriman, CEO of the Chartered Trading Standards Institute, describes the AI Scam Watchlist launch as an important step toward helping people understand criminal tactics and the practical actions they can take to protect themselves and their loved ones. He also argues for cooperation across enforcement, government, industry and consumer organisations so society can gain from innovation while reducing criminals’ opportunities to exploit it.

For online businesses, that cooperation expands the practical fraud perimeter beyond attacks aimed directly at their own systems or customers. AI-assisted campaigns can weaken confidence in the digital channels used for customer acquisition and legitimate transactions, giving security, fraud, product and marketing teams a shared interest in those channels. Fraud management consequently includes the conditions under which customers decide whether a genuine online offer deserves trust.

Key highlights

• AI scams weaken trust in genuine offers: More convincing AI-assisted fraud makes legitimate digital activity harder for consumers to assess. Businesses face customer hesitation and withdrawal alongside the direct costs of fraud.
• AI scales familiar fraud: AI makes shopping, holiday, investment and romance scams more convincing, personalised and scalable while lowering the technical expertise required. Security and fraud teams need to account for higher-volume attacks built on established scam models.
• Consumer uncertainty creates business costs: Nearly half of UK adults struggle to distinguish genuine online offers from AI fakes, while 19% avoid some online activity because of scam concerns. Marketing, product and security teams need to treat digital trust as a shared customer acquisition issue.
• Consumer education strengthens fraud prevention: Guidance on fake websites, voice cloning, deepfake video and other AI techniques can help people assess suspicious approaches. Customer-facing teams can reinforce clear verification steps wherever users encounter offers or requests.
• Coordinated disruption addresses fraud at scale: The UK’s planned Online Crime Centre will combine data from government, law enforcement, financial services, telecoms and technology companies to identify and remove criminal infrastructure. Businesses can support ecosystem-wide disruption through stronger intelligence sharing and cross-industry cooperation.