How do we ensure secure AI models and protect confidential data?
As a company that drives digital transformation as a core competence, the confidentiality and integrity of sensitive data is particularly close to our hearts.
Extensive safety measures
We are aware of our responsibility when it comes to protecting information and assets.
Although OpenAI, the provider of ChatGPT, has made it clear in its terms and conditions that customer data can be used to improve the model, we strongly discourage our customers from sharing confidential information via ChatGPT.
Our top priority is to maintain the integrity of our customers’ data and provide a high level of security. That’s why we at Online Group have taken extensive measures to ensure that our AI models remain secure and confidential information protected:
Anonymization of data
Before we feed data into the AI model, we anonymize it thoroughly. By using pseudonyms instead of context-revealing terms, we ensure that no sensitive information is disclosed.
No chat history and no training for confidential data
We have disabled chat history and confidential data training features in ChatGPT. This ensures that no confidential information is stored or used to improve the model.
In-house copilot
To further strengthen the protection of confidential data, we have developed an in-house copilot. This is based on the secure Azure OpenAI services. This allows our customers to securely use their own data without relying on external AI models.
MS Teams integrated ChatGPT bot
As an additional option, we offer our customers a ChatGPT bot integrated with MS Teams. This is based on the Power Platform and uses the OpenAI API. This enables us to communicate securely while keeping confidential information under the control of the company.
Regular review of AI models
We work closely with OpenAI to ensure that ChatGPT and other AI models meet the highest security standards. OpenAI is already working on a secure business version of ChatGPT tailored to the needs of enterprises
We are aware that the protection of confidential data is an ongoing process. That is why we regularly review our measures and ensure that they meet the constantly changing requirements.
We are committed to maintaining the highest standards of security and providing a safe environment for the exchange of information.
Do you have questions about our AI models?
Network with our “Mister AI” Maciej Jedrzejewski. He regularly reports on trends and facts about AI / AI:
