Miro – Review of AI Features

One of the key questions when implementing sophisticated AI, such as digital twins and integrated AI platforms, is how to integrate them with Business and IS Systems methodologies and tools. I mentioned this during a Webinar for the British Computer Society Leadership Team and was sent an invite to use Miro, a Visual Collaboration Platform. It has experimental AI tools for creating models. There is still quite a lot of work required to iron out issues, however it did demonstrate a number of the future potential dangers of AI. The AI Assistant is used to create models, however it can go much further and answer business questions. This doesn’t work correctly at present. The future danger is that an AI Assistant can be used to answer business questions and formulate business plans, and will give answers with confidence, even though it doesn’t have the capabilities required.

The image shows a very simple sketch were AI is asked which IT Function supports Monitoring and Control. First it gave an answer from its training data, next it made up a spurious answer, finally after prompting it gave the correct answer. The issue is that the AI Assistant doesn’t automatically follow the logic path.