Artificial intelligence: are your data protected?

Before using artificial intelligence in the office, here’s how to check what happens to your files and sensitive data. Imagine a microfinance agent who inputs a complete client file into an artificial intelligence (AI) system. Income, account number, credit history, it’s all there. The tool responds in a few seconds. But one question remains unanswered: what happens to this...

June 24, 2026

Imagine a microfinance agent inputting a client's complete file into an artificial intelligence (AI). Income, account numbers, credit history – everything is there. The tool responds in seconds. But one question remains unanswered: what happens to this information once it's sent? Is it stored? Read by someone? Reused elsewhere?

The same scenario plays out at the hospital, prefecture, or university, with a patient record, student copy, or payslip. This article helps you ask the right questions before using AI at work, even if digital technology isn't your field.

What is it, concretely?

Ensuring that an AI protects your data means making sure that what you entrust to it will not be used, stored, or shared beyond what you have agreed to. Unlike software installed on your computer, an online AI often processes your texts on remote servers, sometimes abroad, and may keep a record of your exchanges. Before sending a sensitive file (client list, salaries, medical records, student records), ask yourself four simple questions: who has access to it, how long it is kept, if the data can leave the country, and if it is used to train the tool. You don't need to be a computer scientist to do this well.

Concrete case: what to do and what not to do

Société Mabika Distribution is an SME in Libreville that sells computer equipment to businesses and administrations. Its sales team, six people, receives dozens of quote requests each week. Client contracts, often multi-page PDFs, contain names, phone numbers, addresses, negotiated amounts, and payment terms.

Tired of manually writing follow-ups and standard replies, the sales manager wants to test a free online AI to summarize contracts and generate emails. She tries the tool with three real cases. The results are fast, almost astonishing. She's considering rolling out the practice to the entire team, without consulting management or checking where the files go once pasted into the interface.

What can be done

Read the tool's privacy policy. Are your files used to improve it?
Replace the names with neutral labels (Client A, Supplier B, Contract 1).
Classify the data: public, internal, confidential, or sensitive.
First test with dummy documents, then get approval from management, legal, or IT.
Write a simple internal charter on authorized tools and prohibited files.

What not to do

Stick contracts, pay slips, or bank statements into an unapproved tool.
To settle for a promise displayed on the website without a verifiable document.
Forget attachments: a PDF or spreadsheet can reveal much more than the visible text.
Give an AI full access to messaging, customer databases, or HR files.
Extend a personal test to the entire team without management approval.

Questions to Ask Before Acting

Do these data allow the identification of a person (name, phone number, photo, ID number, account, grade, medical record)?
Do I really need to send this information to the AI, or can I anonymize it?
Does the provider clearly indicate if my data is used to improve their tool?
How long is the data kept? Can we request its deletion?
Who can access it (supplier, subcontractors, internal administrators, other users)?
Where is the data stored or transferred?
Does my organization and the applicable law permit this use?

UNIVGA Viewpoint

In Gabon and the French-speaking African space, a university, clinic, administration, or SME may be tempted to use AI hosted abroad to save time. The right reflex is not to copy a model from elsewhere, but to check local laws on personal data, access and objection rights, as well as conditions for transfer to other countries. For an administration, this can be done through a simple charter per service. For an SME, through progressive rules: public data authorized, internal data regulated, sensitive data validated beforehand.

For the Virtual University of Gabon (UNIVGA), protecting data is not a barrier to innovation. It is a condition for AI to truly serve training, management, and research. In banking, healthcare, human resources, or administration, the question is the same: who accesses what, and for how long? UNIVGA advocates for responsible adoption: train users, set rules before usage, document choices, and involve business units.

On a continental scale, protecting data also means preserving trust, digital sovereignty, and the credibility of African organizations in the face of imported tools. Good governance does not block innovation. It allows students, teachers, managers, and professionals to use it with peace of mind.

Sources

To go further

17 August 2026

Digital cadastre: how to avoid land errors?

A Colorado example sheds light on land digitization: registering markers online can reduce disputes, if the evidence remains…

Read the article

14 August 2026

How to frame AI at the university without cheating?

Generative AI is disrupting courses and assignments. A simple method to frame it without weakening learning.

Read the article

7 August 2026

Data centers: how to secure the entrance?

Data centers do not only protect themselves with software. Discover the right guarding reflexes to prevent an open door…

Read the article

Artificial intelligence: are your data protected?

Excel with AI: how to maintain reliable results?

Can business software steal your sensitive data?

Can AI enhance the reliability of accident investigations?

Professional Certificate in Governance and CSR — CSR Director

What is it, concretely?

Concrete case: what to do and what not to do

Questions to Ask Before Acting

UNIVGA Viewpoint

Sources

Digital cadastre: how to avoid land errors?

How to frame AI at the university without cheating?

Data centers: how to secure the entrance?

Cookies and Privacy

Set cookies

Professional Certificate in Production & Transformation of Natural Rubber

Shopping cart

Excel with AI: how to maintain reliable results?

Can business software steal your sensitive data?

Can AI enhance the reliability of accident investigations?

Professional Certificate in Governance and CSR — CSR Director

What is it, concretely?

Concrete case: what to do and what not to do

Questions to Ask Before Acting

UNIVGA Viewpoint

Sources

Related Articles.

Digital cadastre: how to avoid land errors?

How to frame AI at the university without cheating?

Data centers: how to secure the entrance?

Shopping cart