Brief IA
›

71% of British Employees Used Unapproved AI

Business & Startups·Tom Levy·

71% of British Employees Used Unapproved AI

71% of British Employees Used Unapproved AI
Key Takeaways
1Microsoft reports that 71% of British employees have used unapproved public AI at work.
2More than a quarter of the affected employees state that they have no approved alternative at their employer.
3An unnamed organization reports having integrated ChatGPT and Claude into an approved professional environment.
Why it matters : Using personal accounts or unapproved services can deprive companies of visibility into the flow of their data.

Microsoft indicates that 71% of British employees have already used an unapproved public AI at work. More than half use these tools every week.

An organization regulates established practices

With a personal account or an unapproved service, a company may no longer know what information is being transmitted, what processing is being applied to it, or where it is going. An organization, whose identity is not specified, reports having integrated ChatGPT and Claude into an approved environment after studying the tasks on which its employees saved time. They previously used these tools to prepare presentations, write, and create other routine content.

This organization claims to have made shared skills accessible to all its employees. In ChatGPT or Claude, these skills can combine elements of institutional discourse, brand vocabulary, and the information necessary for an accurate presentation of the company. They are particularly used to produce slideshows following common guidelines and a brand-compliant presentation.

Employees can thus retain the tools they find useful, accompanied by internal information and shared skills, the organization asserts. It states that it can monitor their usage and the flow of data, with security protections and access rights management.

Personal usage extends to work

Among the employees affected by Microsoft's findings, 41% also use these services in their personal lives. More than a quarter report that their employer does not provide any approved solution. Professional uses include information analysis, content production, or speeding up tasks that would otherwise take hours.

La lettre du soirGratuit · 19 h

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

However, companies from various sectors are already providing approved tools, without their mere availability necessarily changing habits. Public services allow for testing different approaches and alternating between tasks; they create expectations of flexibility, speed, and ease of trial. Employees may expect similar simplicity and utility from professional solutions.

A more restrictive approved option that is less user-friendly or poorly suited to the work may therefore fail to retain its users. Banning other services may mitigate the immediate risk, without making the proposed tool more useful or eliminating the reasons for the initial choice.

Independent controls of the tools used

A professional tool must protect internal information, respect access rights, integrate into existing systems, and meet governance requirements. These obligations apply even as ChatGPT, Claude, or Gemini may currently be preferred by employees. The arrival of new services and improvements to existing tools could alter these choices.

Controls organized at the scale of the AI environment, rather than around a particular model, would allow for changing tools without having to rebuild all protections. They would focus on accessible data, operations permitted for employees, and the visibility necessary for the company. This approach would involve identifying useful services, integrating them, and securing their use, rather than anticipating which one employees will choose next.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.