Your Staff Are Already Using ChatGPT. Is Your Company Data Safe?

Durban: As workplace use of ChatGPT grows, South African businesses need practical rules, employee training and human oversight to protect confidential information while still benefiting from AI.

Across South African workplaces, employees are already using ChatGPT to draft emails, summarise documents, analyse information, prepare reports and solve everyday business problems.

The important management question is no longer whether employees will use generative AI. It is whether they understand what information they should and should not put into it.

This creates a challenge that many organisations have not yet fully addressed. An employee may be trying to save 20 minutes on a task, but in doing so could paste confidential company information, customer details, financial data, HR information or commercially sensitive material into an AI tool without considering where that information is going or whether they are authorised to share it.

The technology may be new. The responsibility for protecting business information is not.

Shadow AI is already a management issue

“Shadow AI” describes employees using AI tools outside formally approved company processes or without adequate governance.

It does not necessarily involve malicious behaviour. In many cases, the employee is simply trying to work faster.

Someone struggling with an Excel formula may ask ChatGPT for help. A manager may paste a long document into an AI tool and request a summary. An employee preparing a proposal may ask ChatGPT to improve the wording. A finance employee may use it to interpret data or help explain a variance.

Each activity may appear harmless until the information being supplied to the AI is considered.

That changes the executive question from “Are our employees using ChatGPT?” to “Do our employees know how to use ChatGPT responsibly with company information?”

For South African organisations, this needs to form part of the broader conversation around confidentiality, information security, internal policies and the Protection of Personal Information Act (POPIA).

Simply banning ChatGPT does not solve the problem

An outright ban may appear to be the safest response, but it does not necessarily prevent employees from using publicly available AI tools.

Employees who believe a tool makes their work easier may continue using it through personal accounts or devices. The organisation can then end up with less visibility rather than more control.

A more practical approach is to establish clear boundaries.

Employees need to understand what information may be used with approved AI tools, what information should never be entered, which systems the organisation permits, when AI-generated information requires verification and who remains accountable for the final work.

This is as much a skills and management issue as it is a technology issue.

Training needs to go beyond prompting

Many early discussions about ChatGPT training concentrated on writing better prompts. That is useful, but workplace adoption requires considerably more.

Employees should understand the difference between public and confidential information. They need to recognise when personal or commercially sensitive information is involved. They should understand that a confident AI response is not necessarily a correct response.

Most importantly, they need a practical validation habit.

College Africa Group (CAG) uses a straightforward principle when teaching employees to work with AI:

Source Data ? Process ? Required Output ? AI ? Validate Against the Source

AI can assist with the work, but employees remain responsible for checking calculations, assumptions, completeness, confidentiality, interpretation and the final business decision. This human-validation principle is central to CAG’s approach to workplace AI. PR Master Prompt

“ChatGPT is already part of the modern workplace. The question is not whether your staff are using it, but whether they are using it safely and productively. Training and clear guidelines give organisations greater control without losing the real business benefits.”

— Arnold Muscat, Director, College Africa Group

Governance and practical skills need to work together

A policy sitting on the company intranet will not, by itself, change employee behaviour.

Businesses need both governance and capability.

Management should establish what tools are approved and what information may be used. Employees then need practical training that applies those rules to the work they actually perform: spreadsheets, emails, reports, documents, research, customer communication and management information.

The objective should not be to frighten employees away from AI. It should be to help them recognise where AI is useful, where the risks lie and when human judgement must take over.

For organisations introducing ChatGPT into everyday work, CAG’s ChatGPT for Business training in South Africa focuses on practical workplace application, responsible use, source validation and human accountability.

The businesses likely to gain the most from ChatGPT will not necessarily be those whose employees use it most frequently. They will be those that combine useful AI tools with sound business processes, stronger employee skills, sensible governance and people who know when to check the answer.

College Africa Group (CAG) has delivered corporate training across South Africa since 2003, helping organisations improve workplace productivity through Excel, Microsoft 365, AI and ChatGPT training, AI consulting and workflow improvement. Better Processes. Stronger Skills. Smarter AI. Contact: sales@collegeafricagroup.com | 083 778 4903