F-Secure's 2026 Digital Trust Report examines what consumers do with AI and finds a widening gap between how widely they adopt the tools and the safety features needed to make those experiences trustworthy, the company said. The findings draw on research with 1,500 consumers in the US and UK, covering how people verify AI answers and what risks they are prepared to take.
The report says people who use AI across many types of tasks are less likely to check its answers consistently than those who limit AI to certain situations. It also found that nearly half of AI users do not question an answer further once it sounds right, even though they know AI can be wrong. Among consumers who do not currently use AI chatbots, scam detection, data protection and the blocking of harmful or manipulative content were among the security features they valued most.
For digital service providers, the report sets out three actions: build on existing customer relationships, design around how people actually use AI, and step in when protection is needed. Dr. Laura James, vice president of research at F-Secure, said trustworthiness is the part a product builder or service provider can act on, and that it rests on being honest about what a system can and cannot do, doing the job well and being reliable, which she described as design and engineering choices rather than marketing ones.
Dr. Amel Bourdoucen, senior user researcher at F-Secure, said people know AI can be wrong but that this does not mean they will always stop and check, and that when an answer sounds right, or there is no clear source to check it against, they often move on. She said AI experiences would be more trustworthy if people were supported in those moments rather than left to carry the responsibility themselves. The report also draws a distinction between using AI and trusting it, noting that consumers are skeptical of the technology yet still rely on its answers, share sensitive information with it and depend on it for high-stakes decisions.
The full report is available now. F-Secure said it will meet executives and experts at Network X in Vienna from October 13 to 15.
