web
You’re offline. This is a read only version of the page.
close
Skip to main content
Community site session details

Community site session details

Session Id :

Power Platform | May the 4th of AI be with you

Carsten Groth mscrm Profile Picture Carsten Groth mscrm 2,085

With all the noise around generative AI and Copilot´s being added to almost any Microsoft product line that most of us are aware of, it´s no suprise that in recent conversations generative and especially responsive AI becomes a top topic. On May 4th, it´s therefore no suprise me running into another conversation around AI becoming either to support or being in need of becoming banned to trasnform people working. Us using artificial intelligence for quite a while now, sometimes without us knowing about it, it becomes obvious that AI becomes casual and success stories around the short-term growth of ChatGPT community will become part of pretty much every industry strategy we´re aware of.

Visual outlining the transition of people working side by side to side by side with AI

I´ve recently finished reading the book „The Age of AI“ by Henry A Kissinger (Author), Eric Schmidt (Author), Daniel Huttenloche and came across this interesting part:

Although AI can draw conclusions, make predictions, and make decisions, it does not possess self-awareness – in other words, the ability to reflect on its role in the world…But inevitably, it will change humans and the environments in which they live.

The Age of AI, Book

This one certainly opens the field of discussion what kind of responsibility is needed in terms of the IT department for instance serving as a service provider and allowing employees to use AI in context of their work. Furhtermore, enabling Fusion Teams inside their development projects to use AI services and even publish those services to end-users in a secured and governed manner.

When it comes to Star Wars (May 4th), we all know that the good and the bad force is ironically fighting for becoming the leader to transform the universe. But in the end it´s guidance and governance that leads to the saga we´re all aware of. When it comes to context of Microsoft AI solutions, it certainly points to running a conversation around responsible AI and how the Power Platform inherits these principles. This concept isn´t new and I´ve been talking about this for quite a while now and even earlier, in October last year, I published an article about this.

Visual outlining Power Platform & Microsoft´s responsible AI principles

While Makers (The Luke Skywalker´s) certainly do have an interest to learn and upskill, and even though first trial might be kind of „let me play and fail or succeed with it“ – there´s certainly a high responsibility on IT- and Security teams (Master Yoda´s) to help govern, secure the company usage of AI services, and obviously train the Makers to responsibly use AI, such as Azure Open AI or Power Platform´s AI Builder. Therefore, starting conversations around responsible AI and upskilling end-users (Luke Skywalker´s) you could use above visual to talk about the principles that should be implemented insight each company. When prompting Large Language Models to gather the insights via AI or generate new data based on it, it should become an obvious task to cross-check results by humans and get rid of possible hallucination.

And while some of this can be pre-scripted, implemented via Governance or DevOps tasks, it´s quite important that every Maker understands the companies intention of using AI services for the good. Microsoft in partnership takes care of offering Azure OpenAI service – so that passing data to 3rd party AI services can be avoided.

To best find a working governance and security concept when using AI services, it is beneficial to understand what kind of business use-cases GPT potentially unlocks. For a rough overview and as a conversation starter, I am sharing the following visual.

Visual outlining typical business scenarios GPT can help unlock

Using these examples, it can be beneficial to define first-class company enabled AI business scenarios that afterwards a Governance & Security + a Guidance can be worked on. Remember it is Master Yoda and Luke Skywalker train together and side-by-side. Prompt engineering is kind of a new skill that we all need to familiarize with rapidly, same as we figured out over the years of using search engines on how to refine our researches. None of us is born as prompt engineer, though we will add this skill over the course of the next years to come, to more efficiently and effectively perform our business tasks.

May the 4th be with you. Until then,…


This was originally posted here.

Comments

*This post is locked for comments