Skip to main content
Cybersecurity

UK Workers Freely Share Passwords with Unmanaged AI Tools

Workers in a brightly-lit office setting, one distracted by a smartphone.

One in five UK employees — 19% — have shared a password, API key or login credential with an AI tool or agent, according to a new Dashlane survey that calls the behavior an emerging “AI Colleague Effect.”

Scope of the Dashlane survey and headline findings

Credential security specialist Dashlane polled 1,000 UK workers and 250 IT executives as part of a wider survey of 3,800 respondents across the UK, US and France for its report Misplaced Trust: Inside the Workplace’s New AI Security Blind Spot. The UK sample showed widespread, often unexamined use of unmanaged AI tools inside companies:

  • 48% of UK workers have allowed AI to perform at least one activity without seeking permission first, including accessing emails, company files and customer information.
  • 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 later discovered it should not have accessed; 45% of those say 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% failed to check where information would be stored or sent, and only 25% say they always consider security risks when using an AI agent at work.

How Dashlane frames the security risks

Dashlane warns that sharing credentials with unmanaged AI and permitting agents to act on accounts creates significant compliance and security risks. The report flags three concrete concerns: information may end up on third‑party servers that could be compromised; data shared with models may be used to train them and later regurgitated in responses to other users; and agents themselves are “a possible target for hijacking by threat actors.”

Security leaders report incidents — yet say they know every agent

The survey found a striking tension between reported incidents and executive confidence. Seventy‑two percent of IT and security executives polled said their organization experienced a security incident involving an AI agent’s credentials or access in the past 12 months. Yet 92% of those same executives claimed they were confident they knew about every AI agent operating within their organization. That sits against a quarter of employees who said they use AI tools their employer does not know about.

Business pressures are also visible: 68% of UK IT security executives said they feel caught between pressure to deploy AI quickly and the need to keep the organization secure.

Dashlane CPO Christophe Frenet’s diagnosis and prescription

Dashlane’s chief product officer, Christophe Frenet, coined the report’s central phrase: “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.”

Frenet framed the problem as a familiar security mismatch applied to a new class of actor. “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,” he said. He urged applying established identity‑security principles to AI: “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.”

What this means for technologists and security teams, policymakers and regulators, and employees

  • Technologists and security teams: they will need to treat AI agents like identities in their access models, watch for active credentials and orphaned agent access (45% of unintended accesses remain active), and respond to the high incidence of agent‑related incidents (72% reported one in the past year).
  • Policymakers and regulators: the possibility that corporate data shared with AI ends up on third‑party servers or is used to train models creates compliance questions about data flows and reuse that regulators will be watching, given the report’s emphasis on third‑party exposure and model training risks.
  • Employees and end users: the behavior gap is clear — substantial numbers are sharing credentials, copying sensitive material into AI tools and not checking storage destinations, while only a quarter always consider security risks — suggesting user education and enforced controls will be necessary to change practice.

The Dashlane data sketches a clear and avoidable disconnect: employees are extending trust to AI agents in ways that security leaders say have already resulted in incidents, even as those leaders express confidence they know every agent in their environment. The report’s prescription is straightforward — treat AI agents as identities and apply the same access and credential hygiene principles already used to protect people — but the survey suggests doing so will require both organizational changes and a halt to casual credential sharing.

Original story