Combing AI with Data Protection and Quality Assurance
Generative AI and Large Language Models are at the heart of the decídalo architecture. This brings enormous benefits but often also raises two main concerns: Data privacy and reliability of the results.
Let’s address both here.
Data Privacy – Yes, We Are GDPR Compliant
Everything we do is GDPR compliant and aligns with the EU AI Act. GDPR (General Data Protection Regulation) and the EU AI Act are sets of rules that all companies doing business in the EU must follow. American regulations tend to be more lenient and differ from the stricter European standards.
Most AI companies are American, which can make working with them tricky, since American data protection standards are less strict than Europe’s.
To stay compliant and simplify the approval process within our customers we selected Microsoft Azure OpenAI as our main AI service. Most of our clients already use Microsoft 365 and Copilot, which use the same services. Microsoft operates the OpenAI language models in their own data centers, including multiple locations within the EU. Our customers can opt to use only EU data centers.
The OpenAI language models are pre-trained (that’s what the P in ChatGPT stands for). Your input is not used to train the models or answer someone else’s questions.
Human in the loop
“But AI makes mistakes.” Yes, but so do humans.
When talking about AI reliability it’s important to first clarify the objectives.
We use AI to automate labor intensive tasks that were previously performed by a back-office person. You can think of decídalo’s AI services as your AI Resource Manager. Instead of sending tasks to your resource management colleague, you give it to decídalo.
Our goal is to complete the task much faster (seconds, not hours or days) and better than the average Resource Manager. We don’t claim to be perfect.
We keep humans in the loop. You can review and adjust all AI generated results. We give you insights into the AI reasoning to make the process transparent.
It is like starting to work with a new coworker. In the beginning you will check his work before you use it. After some time, you know where you can trust him blind and what you should still review.
Have More Questions?
“Move fast and break things” is not the best approach for sensitive data and business continuity. We believe in a future of automation that is secure and transparent.
Feel free to contact us if you want more information or have another question. We understand that different companies have different security and compliance requirements. With over 15 years of experience, we are happy to advise you on data protection and regulatory requirements.