The 76% Figure Comes From a 2025 Online Survey
The timing and sample matter. Lloyds published the generational analysis in August 2026, but the responses were collected between 3 and 28 July 2025. The survey covered 5,000 adults who were already online and able to complete an online questionnaire, with weighting intended to make that digitally engaged group representative by age, gender and region.
The headline should therefore be read as 76% of digitally engaged 18 to 24-year-olds in the survey, not 76% of every young adult in Britain in 2026. The report explicitly excludes people who are offline and says it is not an assessment of digital access or inclusion. The wider 2025 Consumer Digital Index also used face-to-face interviews with 300 digitally disengaged people and other behavioural data, but the AI-use figures came from the online survey.
Lloyds defined personal-finance use broadly. It included budgeting, savings planning, financial education, insurance comparisons, investment research, debt strategies, mortgage calculations and future planning. Using a chatbot to explain compound interest therefore sits in the same overall adoption measure as seeking mortgage or investment information. The figure does not mean three in four young adults handed decisions to an autonomous system or relied on AI as regulated advice.
Young Adults Lead Adoption but Not Confidence
The result continues a pattern found in other UK research. An earlier 5,000-person survey reported that 67% of adults aged 25 to 34 and 53% of those aged 21 to 24 used AI platforms for financial guidance. That study also found that some younger respondents would consult social media before a bank or professional adviser.
Lloyds’ data adds a useful limit to that adoption story. Among people using digital tools for money management, confidence rose from 64% at ages 18 to 24 to 85% at ages 65 to 74. The age difference may be connected to experience, but the survey cannot isolate that explanation from income, assets, employment, debt, housing, education or the type of financial decision each person faces.
A 22-year-old may be using AI precisely because budgeting, credit and investing are unfamiliar. A 68-year-old may have decades of experience, a paid-off mortgage and established financial routines before opening a digital tool. Lower confidence could lead to higher AI use rather than result from it, while both could be driven by the same stage-of-life pressures.
Jas Singh, CEO of Consumer Relationships at Lloyds Banking Group, said “confidence with money is something that’s built over time.” He argued that the useful combination is digital capability plus financial knowledge, with younger users bringing familiarity with technology and older users contributing experience.
AI Use Changes as Financial Needs Change
Among respondents who had used AI for personal finance, 64% of those aged 25 to 34 used it for budgeting and planning, the highest share in any age group. For people aged 35 to 44, 28% had used it for mortgage advice. Among 55 to 64-year-old AI users, 57% used it to compare insurance products.
These differences are consistent with the financial tasks that become more relevant at different life stages. Younger workers are more likely to be establishing budgets and savings routines. Mortgage decisions become more prominent during household formation, while insurance selection and retirement planning gain weight later. Age may therefore influence what question is asked, while the usefulness of the answer depends on the data and context supplied.
Financial institutions are beginning to embed that context rather than leaving users with a blank general-purpose chatbot. Starling Bank’s agentic assistant can analyse transaction history, organize bills and create savings plans inside the banking environment. In investing, Robinhood’s UK AI digests combine news, research, technical indicators and account-level access to explain stock moves in plain language.
The advantage of an embedded tool is access to verified balances, transactions and product information, subject to the provider’s controls. A general chatbot typically depends on whatever the user types and may not know current interest rates, tax status, debt terms or risk tolerance. Embedded access can improve relevance but also increases the importance of privacy, security, consent and the provider’s responsibility for the output.
Use Is Not the Same as Trust
Lloyds’ main 2025 index found that 56% of digitally engaged adults had used AI for personal finance during the previous year and almost one in three used it weekly. ChatGPT was the most-used tool among AI users at 60%, followed by bank AI assistants at 32%. Dedicated financial AI apps were used by 9%.
Most users still expressed reservations. Eighty-three percent were concerned about data privacy, 80% about inaccurate or outdated information, 72% about bias and 69% about insufficient personalization. The annual saving attributed to AI averaged £399 among users, but Lloyds described that number as an unverified perception rather than an outcome confirmed through transactions.
The tension is also visible in investing. Research published earlier this year found that 46% of UK adults trusted ChatGPT for crypto education, while 54% of those who had used AI for crypto research went on to invest. That sequence does not prove that AI caused the purchase, but it shows why the boundary between education, recommendation and action matters.
General AI Does Not Carry Advice Protections
The UK Financial Conduct Authority warns that information from general-purpose tools such as ChatGPT and Copilot is not regulated financial advice. Its consumer guidance on AI investment research says users generally cannot rely on the Financial Ombudsman Service or Financial Services Compensation Scheme for harm caused by unregulated AI information. Protections may apply where an authorized firm uses its own AI system to provide a regulated service.
This distinction becomes more important as systems move from answering questions to acting. An FCA survey of more than 5,000 retail-finance consumers found that one in five would consider an AI agent able to act within preset goals. The regulator’s subsequent review warned that agentic systems could change saving, investing, switching and payments while increasing questions about liability, fraud, conflicts and control. The wider implications are examined in the FCA’s review of AI’s effect on retail finance through 2030.
The Pensions Regulator has made a similar distinction for long-term planning. Its 2026 AI plan says general-purpose tools are unregulated and identifies inaccurate information, liability, bias and scams as risks. An error affecting a pension can compound for years, making confident presentation a poor substitute for verified assumptions and authorized advice.
The Commercial Opportunity Is Guided Use, Not More Prompts
Lloyds’ findings point to a design problem for banks and fintech firms. Younger adults do not appear to need encouragement to try AI. They need help judging when an output is educational, when it is a product suggestion, what information is missing and when a decision requires regulated or human support.
Older users present a different opportunity. Adoption is lower, but confidence among digitally engaged people is higher and use is concentrated in defined tasks such as insurance comparisons. That may favor tools with a narrower purpose, cited data, clear limitations and an easy route to a person rather than open-ended conversational systems.
The UK advice market is moving in that direction. From April 2026, firms with the appropriate permission can provide targeted support to groups of consumers with common characteristics, while the FCA is also considering simpler forms of individual advice. Banks may be able to combine AI-led identification of needs with services operating inside a clearer regulatory boundary, rather than allowing a general chatbot to appear more authoritative than it is.
The headline adoption number is still significant: AI has entered everyday budgeting, comparison and research for many digitally engaged young adults. The deeper result is that access has run ahead of confidence. Closing that gap will require financial education, transparent sources, current data and escalation to protected advice, not simply a more persuasive answer generated on demand.