The Hidden Risks Behind Letting ChatGPT Shop With Your Credit Card

The future of online shopping may have arrived sooner than many expected, but not everyone is convinced it is a future worth celebrating.

Visa's new partnership with OpenAI will allow ChatGPT to move beyond recommending products and into something far more significant. The artificial intelligence platform will soon be capable of independently completing purchases on behalf of users. Supporters say the technology could transform ecommerce by making online shopping faster and more convenient. Critics, however, warn that the move could create serious security, fraud, legal and financial challenges for consumers and businesses alike.

For shoppers in Newcastle and across the UK, the announcement raises an uncomfortable question. If an AI system makes a purchase you never intended, who is ultimately responsible?

That question currently has no clear answer.

A New Era Of AI Shopping Begins.

The partnership will allow consumers to connect Visa payment credentials directly to ChatGPT. Instead of manually browsing websites, comparing products and entering payment details, users can instruct the chatbot to find items, compare prices and complete transactions automatically.

In theory, the experience sounds simple.

A user could ask ChatGPT to locate wireless headphones under a specific budget, compare retailers and complete the purchase within seconds. The process removes several traditional steps from online shopping and promises a more streamlined customer experience.

However, convenience often comes at a cost.

Cybersecurity specialists have spent years warning consumers about the dangers associated with online fraud, identity theft and compromised payment information. Introducing autonomous purchasing agents into an already complex ecommerce environment could potentially create entirely new vulnerabilities.

Many experts believe the technology is arriving before regulators, retailers and payment providers have fully addressed the risks.

Why Security Experts Are Concerned.

Online fraud is already one of the fastest-growing forms of crime in the UK.

According to UK Finance, fraud losses reached £1.17 billion during 2024, while more than 3.13 million cases of unauthorised fraud were reported, representing a significant increase compared with the previous year.

Remote purchase fraud, often referred to as card-not-present fraud, remains one of the largest threats facing online shoppers. These crimes occur when stolen payment card information is used to complete transactions online without the cardholder's knowledge.

Industry data shows card-not-present fraud accounts for around 70 percent of UK card fraud losses and continues to rise despite improvements in authentication systems.

Security researchers fear AI shopping agents could inadvertently increase opportunities for criminals.

If hackers gain access to an AI assistant linked to payment information, the consequences could extend far beyond a single fraudulent transaction. A compromised AI agent may be capable of making multiple purchases across several retailers in a matter of minutes.

Unlike traditional fraud, where criminals must manually enter payment information, automated AI systems could dramatically increase the speed and scale of attacks.

For Newcastle residents already dealing with rising levels of digital fraud, the prospect is understandably concerning.

The Data Breach Problem Nobody Has Solved.

Another major concern revolves around data protection.

AI shopping agents require access to significant amounts of personal information in order to function effectively. This can include payment details, shopping preferences, delivery addresses, purchase history and behavioural data.

The more information connected to an AI system, the more attractive it becomes to cybercriminals.

Recent years have seen a steady stream of major data breaches affecting global corporations. Every breach creates new opportunities for fraudsters to gather information that can later be used for identity theft and financial crime.

Experts warn that linking payment systems directly to AI platforms increases the amount of sensitive information concentrated in a single environment.

Even if payment credentials are tokenised or encrypted, attackers often target the weakest link in a system rather than the strongest.

For consumers, the result could be unauthorised purchases, account takeovers and lengthy disputes with banks and retailers.

What Happens If ChatGPT Buys The Wrong Product?

One of the biggest unanswered questions involves liability.

If a consumer instructs ChatGPT to find a laptop under £1,000 and the AI accidentally purchases the wrong model, who bears responsibility?

If the product arrives exactly as ordered by the AI but not as intended by the customer, the situation quickly becomes legally complicated.

Industry leaders speaking at technology conferences have acknowledged that responsibility for AI-generated purchases remains a largely unresolved issue.

Some companies may attempt to transfer responsibility to users through terms and conditions. Others may place liability on merchants, payment providers or software developers.

The reality is that courts have yet to establish clear legal precedents.

This uncertainty could become particularly problematic as AI purchasing expands into higher-value sectors such as travel, electronics and financial services.

Merchants Could Face A Chargeback Surge.

Businesses may have even more to worry about than consumers.

Chargebacks already represent a major challenge for online retailers. These disputes occur when customers ask their card provider to reverse a transaction, often claiming fraud, non-delivery or unauthorised activity.

The introduction of AI shopping agents could create a new category of dispute.

Consumers may argue they never intended for an AI assistant to complete a specific transaction. Retailers, meanwhile, may insist the purchase was authorised because the AI acted within permissions granted by the customer.

The result could be a dramatic increase in contested transactions.

Research cited by Mastercard estimates businesses worldwide could lose approximately $15 billion due to fraudulent chargebacks during 2025, highlighting the growing financial burden merchants already face.

For small businesses in Newcastle's retail sector, even a modest increase in chargebacks could significantly impact profitability.

Many independent retailers operate on tight margins. Repeated disputes can lead to lost revenue, additional processing fees and higher payment costs.

Fraudsters Are Already Using AI.

The timing of Visa's announcement comes as fraud investigators report growing misuse of artificial intelligence by criminal groups.

Recent fraud reports have highlighted increasing use of AI-generated scams, synthetic identities and sophisticated impersonation techniques.

Cybercriminals are leveraging AI to create convincing phishing emails, cloned voices and fake customer service interactions.

Some experts fear autonomous shopping agents may become another tool exploited by organised crime networks.

A criminal who gains access to stolen credentials could potentially create multiple AI agents capable of making purchases automatically across numerous websites.

This could overwhelm fraud detection systems that were originally designed to identify human purchasing patterns rather than machine-driven activity.

The challenge facing payment providers is immense.

They must distinguish between legitimate AI-assisted purchases and fraudulent automated transactions without creating excessive friction for genuine customers.

Newcastle Businesses Watching Closely.

The implications are particularly relevant for businesses across Newcastle and the wider North East.

The region has experienced significant ecommerce growth over the past decade, with many local retailers increasingly dependent on online sales channels.

While AI-driven commerce could create new revenue opportunities, it may also introduce additional operational risks.

Merchants could face increased fraud monitoring costs, more chargeback disputes and higher compliance requirements.

Some business owners are also concerned about the growing influence of large technology platforms in the purchasing process.

Traditionally, retailers controlled much of the customer journey. AI shopping agents may shift that power towards technology companies that act as intermediaries between buyers and sellers.

This could reduce direct customer engagement and make businesses more dependent on external platforms.

The Trust Problem.

Perhaps the greatest obstacle facing AI shopping is trust.

Consumers have become comfortable using artificial intelligence to answer questions, generate content and provide recommendations. Allowing an AI system to spend money is an entirely different proposition.

Financial decisions carry emotional weight.

Even relatively small purchases can create frustration when mistakes occur.

Research into payment fraud consistently shows that victims often suffer psychological stress regardless of the financial amount involved.

When an AI system is given authority to make purchases, consumers may struggle to determine whether an error resulted from fraud, software malfunction or their own instructions.

That uncertainty could undermine confidence in the technology.

Many analysts believe widespread adoption will only occur once consumers are convinced robust protections exist.

Regulation Is Struggling To Keep Up.

One recurring theme among technology experts is that innovation often moves faster than regulation.

AI shopping appears to be no exception.

Legal frameworks governing autonomous transactions remain fragmented and incomplete.

Questions surrounding consent, liability, consumer rights, refunds and merchant obligations continue to be debated across the industry.

Some specialists believe governments may eventually need to introduce dedicated regulations covering AI-powered commerce.

Without clear standards, disputes could become increasingly common.

Consumers may find themselves caught between software providers, payment companies and retailers, each arguing that responsibility lies elsewhere.

Until regulators provide greater clarity, uncertainty will remain a significant barrier to trust.

The Future Could Be Convenient But Risky.

There is little doubt that AI-powered commerce offers genuine benefits.

Consumers could save time, discover better deals and automate routine purchases.

Businesses may gain access to new customers and more efficient sales channels.

Yet the technology also introduces risks that cannot be ignored.

Fraud, identity theft, unauthorised purchases, data breaches and chargeback disputes already cost billions annually. Adding autonomous purchasing agents to the mix could amplify those challenges before solutions are fully established.

Statistics from UK Finance show that fraud continues to represent one of the most significant threats facing the UK's financial ecosystem, with losses exceeding £1 billion annually and millions of fraud cases recorded every year.

Against that backdrop, consumers in Newcastle and beyond may be justified in approaching AI shopping with caution.

The technology may eventually become as normal as contactless payments are today.

For now, however, many questions remain unanswered.

And when artificial intelligence is spending real money, unanswered questions can quickly become expensive mistakes.

Share your thoughts.

Do you trust AI to make purchases using your bank card, or do the fraud and security risks outweigh the convenience?