Large numbers of UK employees are sharing sensitive information with unmanaged AI tools and allowing them to take actions on their accounts without considering the security risks, Dashlane has warned.
The credential security specialist polled 1000 UK workers and 250 IT execs as part of a wider survey group of 3800 across the UK, US and France to create its latest report, Misplaced Trust: Inside the Workplace’s New AI Security Blind Spot.
It found that a fifth (19%) have shared a password, API key or login credential with an AI tool or agent, while nearly half (48%) have allowed AI to perform at least one activity without seeking permission first, including accessing emails, company files and customer information.
Read more on AI risks: NCSC Publishes Guidance on Securing Agentic AI Use
Dashlane unearthed a high propensity among UK workers to use shadow AI without thinking about the repercussions for their employer.
It found that:
- 37% would be willing to lend their login credentials to a new AI tool after one week
- 15% have granted an AI tool access to a file, system or account they subsequently discovered it should not have accessed, with 45% claiming the access remains active
- 23% have copied sensitive information into AI that they would not normally share with a colleague, website or other work system
- 21% have failed to check where information would be stored or sent
- Just a quarter (25%) say they always consider security risks when using an AI agent at work
“Our latest data shows an emerging ‘AI Colleague Effect’ where employees are overextending trust to AI systems, sharing credentials and granting access to AI agents, sometimes without fully understanding what those tools can do with the permissions they’ve been given,” explained Dashlane CPO, Christophe Frenet.
“That creates significant challenges for security teams trying to detect misuse or investigate incidents.”
Security Leaders Are Over-Confident
Sharing sensitive information with LLMs/agents or allowing agents access to internal corporate systems represent a significant compliance and security risk.
Information may end up on third-party servers which could be compromised, and it may be used to train models and regurgitated in responses to other users. Agents are a possible target for hijacking by threat actors.
In fact, 72% of IT and security executives Dashlane polled said their organization has experienced a security incident involving an AI agent’s credentials or access in the past 12 months.
Yet the vast majority (92%) also claimed they were confident that they know about every AI agent operating within their organization, the report found. That contradicts the claims from a quarter (25%) of employees who said they use AI tools their employer does not know about.
Business demands for adoption are weighing heavily on the minds of security leaders.
More than two-thirds (68%) of UK IT security executives claimed to feel caught between pressure to deploy AI quickly and the need to keep the organization secure.
“The challenge for security leaders is that the pace of change presented by AI and the potential consequences of misuse is unlike any other technology before it,” concluded Frenet.
“The principles that make up the foundation of a good security program for securing human identities – just-in-time provisioning, least privilege, good credential hygiene and authentication – have to be applied to our new AI colleagues as well.”
Leave a comment