SecurityBrief UK - Technology news for CISOs & cybersecurity decision-makers
United Kingdom
Starling launches AI scam checker for five million

Starling launches AI scam checker for five million

Thu, 25th Jun 2026 (Today)
Karen Joy Bacudo
KAREN JOY BACUDO Finance Editor

Starling has launched an AI scam-detection feature in its banking app for the bank's five million customers.

Built into Starling Assistant, the tool is designed to identify romance scams and more than 10 other types of fraud, including investment, pension and deepfake phishing scams. Customers can use voice or natural-language prompts to describe a payment or interaction. The assistant then responds with questions to test whether the request shows signs of fraud.

The launch comes as authorised push payment fraud continues to rise in the UK. UK Finance figures cited by Starling show Britons lost £576.4 million to this type of fraud last year, up 19%. Romance fraud losses rose 23%, while investment fraud losses climbed 40%.

The updated tool builds on a scam-checking service Starling introduced in 2025 for online marketplace listings. According to the bank, that service, in the app's payments section, increased the rate at which customers cancelled marketplace payments by 300%.

How it works

In an example provided by Starling, a customer tells the assistant they have been asked to transfer £3,000 for a plane ticket for a new partner. The system then asks follow-up questions about why the partner cannot pay themselves, where the pair met and how long they have been in a relationship. It then advises whether the situation is likely to be a scam and suggests contacting the bank's support team.

The feature was informed by fraud advocate Cecilie Fjellhøy, whose experience of a romance scam became widely known through the documentary The Tinder Swindler. Fjellhøy lost nearly £200,000 after a scammer persuaded her to take out loans and credit cards in her own name.

Her involvement reflects a wider focus on the psychology of fraud, particularly in romance scams, where victims can become isolated from people trying to intervene. The aim is not only to flag suspicious payments but also to disrupt the pattern of persuasion before money is sent.

"When friends and family try to break that spell, it can cause huge rifts that push you away from the people who genuinely care and even closer to the scammer," said Cecilie Fjellhøy, fraud advocate.

"Sometimes it needs someone or something objective, like your bank, to help you see the scammer's behaviours for what they really are. It's great to have been involved in helping bring this feature to life and see it make a difference to people's lives," Fjellhøy added.

Fraud response

The expansion has also drawn support from the government as ministers seek tougher responses to online fraud. Lord Hanson, Minister for Fraud, linked the announcement to a broader campaign to reduce losses and disrupt criminal networks.

"Fraudsters don't care about the devastating impact romance fraud can have on people's lives. I welcome the expansion of Starling's AI tool and urge everyone to Stop! Think Fraud to protect themselves and those close to them," Lord Hanson said.

"We're investing £250 million as part of our new Fraud Strategy to crack down on this criminality. We will use every tool at our disposal to dismantle criminal networks, bring fraudsters to justice and strengthen protections for victims," he added.

For Starling, the launch extends a broader push to use artificial intelligence in customer-facing banking services. Starling Assistant was introduced in March as a financial assistant that lets customers ask questions about spending, budgeting and transfers through conversational prompts.

The scam-detection function is now integrated into the assistant rather than sitting only alongside payments. The system runs on Google Cloud's Gemini model, with access controlled through customer opt-in. Customer data remains within Starling's Google Cloud environment and is not used to train the model, according to the bank.

Bernadette Smith, Chief Customer and Banking Officer at Starling, said banks still face limits when most scams begin elsewhere online.

"Fraud must be stopped at source, but without effective collaboration from the social media platforms where most scams originate, it's up to individuals and their banks to keep the scammers at bay. At Starling we're working behind the scenes to identify threats and we're empowering our customers with tools to protect themselves," she said.

Harriet Rees, the bank's Chief Information Officer and the UK government's AI champion for financial services, said the tool is intended to change how people think about fraud risk before making payments.

"We're using AI to create behaviour change," said Harriet Rees, Chief Information Officer at Starling.

"We want anyone and everyone to realise that they too could be a victim of a scam. That starts with asking simple questions and being inquisitive about what you plan to spend," Rees said.