IDA uses AI to enhance the member experience and ensure you as a member receive fast and accurate help. Read about how IDA uses generative AI and trains LLMs, IDA naturally takes responsibility for protecting members' data from misuse and data breaches.
When IDA uses AI in member advice, for example in connection with a legal case, you can choose AI-generated answers.
If you accept that AI can contribute to processing your legal case, a specially developed AI agent will read your query, check IDA's sources and, if possible, generate an answer to your query.
You will then be presented with the answer on your legal case.
If the answer from the AI agent is sufficient, you can let the AI agent close your case.
You can bypass the AI agent's answer at any time and get help from a legal councellor instead.
Your data, which is used to generate the AI agent's answer, will be recorded on your legal case in IDA's associated system, and on IDA's AI platform for 30 days for validation.
IDA continuously checks that the AI agent gives you as a member the correct answer, and we continuously update IDA's AI agents with the latest guidelines and legislation for the area in which the AI agent operates.
IDA has developed its own AI platform based on AutoGen Core, Azure OpenAI and Azure AI Search.
IDA's AI agents are tested and developed in collaboration with domain experts for the given area.
This means that when using AI within the legal field, a number of our legal experts will have qualified and tested the AI agent.
All IDA's AI agents are given a number of limitations to eliminate hallucination (misinformation) and use, as far as possible, RAG (Retrieval-Augmented Generation) where IDA's own sources are included in the AI's generation of answers. This ensures better answers to IDA's members with sensible source references to the origin of the answer.
Since IDA uses its own platform and this uses Azure OpenAI, there are a number of measures to secure member data against data breaches.
When using IDA's AI solutions, member data will therefore NOT:
In addition, the following applies:
"The Azure OpenAI Service is operated by Microsoft as an Azure service; Microsoft hosts the OpenAI models in Microsoft's Azure environment and the Service does NOT interact with any services operated by OpenAI (e.g. ChatGPT, or the OpenAI API)." Data is stored on IDA's European Azure OpenAI source on our own Azure tenant. Internally, Azure OpenAI always uses Microsoft's AES-256 encryption as standard, and IDA can delete given member data at any time.