{
  "id": 20741064,
  "title": "'The risks of AI spread faster': Report finds workers are handing over logins and even passwords to AI to make their jobs easier",
  "url": "https://urgent.news/2026/11/19/the-risks-of-ai-spread-faster-report-finds-workers-are-handing-over",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-19T11:05:00.000Z",
  "source": {
    "name": "TechRadar",
    "slug": "techradar",
    "url": "https://www.techradar.com/pro/security/the-risks-of-ai-spread-faster-report-finds-workers-are-handing-over-logins-and-passwords-to-ai-as-security-teams-struggle-to-keep-pace"
  },
  "original_language": "en",
  "account": "A new study by Dashlane reveals a troubling trend among workers who are willingly sharing their login credentials with artificial intelligence (AI) tools to streamline their job tasks. This alarming practice is taking place despite mounting concerns over AI-enhanced cyberattacks.\n\nAccording to the research, nearly one in five UK employees have shared sensitive passwords, API keys, or other login credentials with AI tools, potentially granting the AI access to sensitive business systems. Moreover, over half (48%) of the surveyed workers have allowed AI to execute at least one task without prior approval, which could include handling emails, company files, and customer information. Astonishingly, one in ten (10%) workers have even allowed AI to process payments.\n\nThe report also found that more than a third (37%) of UK workers would be willing to grant login credentials to a new AI tool after using it for just a week. Furthermore, 45% of respondents say they would comply with an AI tool's request to confirm their credentials daily, thereby exposing themselves to potential credential theft or impersonation.\n\nOverall, 44% of workers have utilized AI for work purposes without obtaining prior approval in the past year, while 46% began doing so even before workplace rules were in place. Notably, only a quarter (25%) of employees consistently consider security risks when using AI agents at work, including 15% who have inadvertently given an AI tool access to files, systems, or accounts that should not have been granted access. Nearly half (43%) of those surveyed admit they believe such unauthorized access is still active.\n\nPerhaps most alarming is the fact that nearly one in four (24%) workers have pasted confidential or proprietary information into AI tools that they would not normally share with a colleague or another work system. Furthermore, 42% of the respondents claim they have never received any guidance on what information is safe to share with AI.\n\nDespite these concerning findings, security leaders appear more confident in their knowledge of AI tools within their organizations. While 92% of UK IT security executives claim they are aware of every AI agent in their organization, a quarter of employees admit to using tools their employers are unaware of. The pressure to rapidly deploy AI technologies is likely contributing to this discrepancy, as more than two-thirds (68%) of security executives express feeling caught between accelerating AI adoption and safeguarding their organization.\n\nSome security executives report that deployment moved faster with less security review, with 54% of those facing pressure admitting that the deployment occurred at a quicker pace with fewer security checks. Christophe Frenet, Dashlane's chief product officer, warns that \"the risks of AI spread faster, appear inside already approved tools, and unfold often without any further action from employees.\" He emphasizes the heightened stakes for organizations when employees grant access to AI tools, as these tools may be trusted more than human coworkers.",
  "summary": "Handing over logins to AI may help speed up processes, but could leave your business at risk, Dashlane warns.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}