AI: Nearly One in Five British Employees Shares Their Access Credentials

According to Dashlane, nearly one in five British employees reports having shared a password, an API key, or another identifier with an AI tool. This information could grant access to sensitive company systems.
Executives are already reporting incidents
In the UK, 72% of security executives surveyed by Dashlane say their organization has experienced an incident involving identifiers or AI agent access in the past 12 months. One-third report that an agent gained unauthorized access to data, systems, or identifiers. For 32%, an agent executed an action that was not authorized.
The visibility claimed by these executives contrasts with reported usage: 92% assert they are aware of all AI agents present in their organization. Yet, a quarter of employees say they use tools that their employer is unaware of.
On the employee side, 15% admit to having granted an AI tool access to a file, system, or account before realizing that this permission should not have been given. Nearly 43% believe that the mistakenly granted access has not been revoked, according to Dashlane.
Employees delegate even payments
Among the workers surveyed by Dashlane, 48% report having allowed AI to perform a task without prior validation. This includes emails, professional files, or customer information. One in ten entrusts it with processing payments.
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisissez votre rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
The stated intentions also pertain to access: 37% of British workers would be willing to provide their identifiers to a new tool after a week of use. If an AI requested confirmation of identifiers every day, 45% say they would comply. Dashlane associates this behavior with a risk of credential theft or identity fraud.
Nearly a quarter of workers also acknowledge having copied confidential or proprietary information into an AI that they would not typically share with a colleague or another professional system. Christophe Frenet, product director at Dashlane, sees in these responses a sometimes greater trust in AI colleagues than in human colleagues.
Security guidelines and pressure on deployments
Nearly 42% of workers report never having been informed about the data they can safely share with AI. Only 25% say they systematically assess security risks when using an agent at work. Dashlane sees this as a lack of awareness.
Over the past year, 44% report having used AI in their work without prior authorization; 46% say they started before any rules existed. The pressure to deploy could contribute to these difficulties: 68% of security executives say they are torn between accelerating adoption and protecting their organization. Among those who feel this tension, 54% report accelerated deployments accompanied by fewer security reviews.
For Christophe Frenet, risks are advancing faster, manifesting within already authorized tools, and often occurring without further employee intervention. He believes this increases risks for organizations when their employees grant permissions to AI tools or expose them to sensitive data.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.