How to use Generative AI without compromising your privacy

Main risks of Generative AI

Although these platforms do not have malicious intent, misuse can expose sensitive data. The main risks include:

• Exposing personal or confidential data: By entering sensitive information into a cloud tool, it could be temporarily stored.
• Accidental leaks: If you share conversations containing private data, you might be exposing more than you realize.
• Dependence on external tools: Excessively relying on AI solutions without checking results can lead to errors or bad practices.

Moreover, many of these tools are constantly updated and store conversations to train their models, which can increase risk if permissions and privacy settings are not properly managed.

Here are some data points we recommend not using to avoid leaks.

Data you should avoid entering

• Personally identifiable information (ID numbers, addresses, phone numbers, full names of clients or employees).
• Financial or contractual information.
• Legal or sensitive documents without anonymizing.
• Access keys, passwords, or authentication tokens.

It is also not advisable to share business strategies, prototypes, or any other company content if it is not legally protected. Sometimes, a simple description can contain more information than it seems.

Good security and privacy practices

To use generative AI safely, follow these recommendations:

1. Anonymise information: If you need to share a real case, remove names, numerical data, or specific references.
2. Do not use AI to store information: It is a tool for consultation, not a secure repository.
3. Read and understand the terms of use: Make sure you know what data can be stored or analyzed.
4. Avoid blind automation: Always review generated content before using or publishing it.
5. Train your team: If you work in an SME, all employees should know these guidelines to avoid human errors.
6. Enable security notifications: If your AI tool has alerts about accesses or configuration changes, activate them for greater control.
7. Use complementary tools: If you work with particularly sensitive data, consider combining generative AI with privacy management platforms or test environments.

Practical Example: How to Configure ChatGPT to Improve Your Privacy

• If you use ChatGPT, you can adjust your privacy settings by following these steps:
• Go to Settings > Data and Privacy.
• Disable the option "Use your conversation content for training" if available.
• Delete conversation history you no longer need.
• Periodically review your account settings to adapt them to your needs.

Remember to log out of shared devices and use two-step verification if available, especially if accessing from public networks or shared equipment.

Generative AI is a powerful ally to improve your business's efficiency and creativity.

Adopting good privacy practices is essential to protect your information and avoid unnecessary risks. With knowledge and caution, you can harness its potential without compromising your security.

If you are looking for more information about Artificial Intelligence, you can find more of our AI-related content. Access now!