OPTIVANTAAgency

AI

Using ChatGPT in Your Company: Data Protection, Rules and Good Practice

Many employees already use AI, often without rules. How to use ChatGPT and other assistants securely and in line with data protection.

Melanie Dincdemir · Published on · 7 min read

Key points at a glance

  • Private AI accounts are unsuitable for company data; business versions offer contracts and data protection settings.
  • A short AI policy defines which data may go into AI tools.
  • Under the EU AI Act, companies must ensure sufficient AI literacy among their staff.
  • Results are always checked and owned by people.

Why rules matter now

In many companies, employees already use AI assistants in everyday work, often with private accounts. Confidential information and personal data can end up with providers uncontrolled. Clear rules protect the company and give the team certainty.

Private or business version?

Business versions of AI assistants generally offer a data processing agreement, management of user accounts and settings so that inputs are not used to train the models. They are therefore the better choice for working with company data. Check the terms of the respective provider.

What belongs in an AI policy?

  • Which tools may be used, with business accounts.
  • Which data must not be entered, such as health data, passwords or confidential contracts.
  • How personal data of customers and employees is handled.
  • That results are checked before they are used or published.
  • When AI-generated content is labelled.
  • Who helps with questions and how training works.

What the EU AI Act requires

The EU AI Act obliges companies that deploy AI systems to ensure a sufficient level of AI literacy among their staff. There are also transparency obligations, for example when people communicate with an AI system or when AI-generated content looks deceptively real. Stricter rules apply to high-risk applications.

Good practice in everyday work

  1. Anonymise data before entering it or use placeholders.
  2. Check facts, numbers and sources.
  3. Share good instructions as templates within the team.
  4. Read and adapt customer communication before sending it.
  5. Train regularly and exchange experiences.

This overview does not replace legal advice.

Conclusion

ChatGPT and other assistants can be used securely in companies when business versions, a clear policy and regular training come together. That way you benefit from AI without risking data protection and confidentiality.

Frequently asked questions

Melanie Dincdemir (AI-generated)AI

Melanie Dincdemir

Marketing, digital business and business model development at Optivanta

This article is for general information only and does not constitute legal, tax or investment advice. It does not replace advice on your individual case. Despite careful research, we accept no liability for accuracy, completeness or timeliness; use is at your own risk. Articles bearing an author's name reflect that author's opinion. The provider of and party responsible for all content is Optivanta Solutions LLC. More in the imprint

AI used securely in your company?

We develop AI agents to your rules, hosted locally on your own infrastructure if you wish.