Introduction
Generative AI. You have attended the webinars. Done your investigations . Started a few trials. Written a few prompts. Yeah! You might think you are ready to get going with some real AI. But if you do not take steps to secure your AI, you could be creating some trouble for yourself. Expensive trouble.
What can self-driving cars teach us about responsible AI?
Step back and take a lesson from self-driving cars. The idea of letting your self-driving car be your chauffeur sounded cool at first. Then some high-profile accidents started to occur, and people were less enthusiastic about the technology.
Just recently, Waymo recalled its self-driving cars after accidents in Phoenix. Tesla has issued a massive recall. Chevy Cruise vehicles took a hit (sic) after a report of hitting a pedestrian. And some ticked-off San Franciscans set a Waymo vehicle on fire.
Here is the parallel. Folks jumped onto the idea of self-driving cars, relying on the technology to behave in a responsible and appropriate way. Self-driving car software showed that it hasn’t earned the level of trust that people wanted.
Copilot for Microsoft 365 – Unveiling the Dynamics and Capabilities
Microsoft 365 Copilot is coming soon but is your organization ready? As organizations increasingly embrace Microsoft 365 Copilot for enhanced collaboration and productivity, the strategic planning of its rollout becomes critical.
Get the eBookSecure your AI or pay the price
Generative AI can seem like self-driving cars in a way. You see the power of the technology. But you are not quite sure you know how the technology is going to play out. In the background, you have data privacy regulations that are ready and willing to levy some big fines if your AI application exposes private data. Amazon, WhatsApp, Google, and Meta/Facebook have been handed multi-million-dollar fines for software practices that exposed private data.
Generative AI will get better over time, no doubt about it. Self-driving automotive technology will improve, too. No matter how improved generative AI becomes, you will still want to plan for a worst-case scenario where it accesses data it should not be seeing.
Here is where Microsoft Purview comes into the picture.
Find the Data. Label the Data. Protect the Data. Audit the Data.
Right now, I am managing six projects to find and separate new Company B’s data from divesting Company A’s data. In every one of those projects, we have been challenged to develop a complete inventory of company data assets. There is no authoritative, up-to-date record of where data assets live. So, we do our best to find the data and make plans for the eventual “surprise!” moments when new data turns up.
Your company’s data is everywhere. Purview helps you find your data. It tells you what you have, where you have it and how well it is protected.
With Purview, you can create labels for different kinds of data and then associate those labels with security policies.
Next, you can audit your data, to track when the data is accessed and changed.
Finally, you can assess how your data protection compares against industry privacy and security standards. Do you want to hope you are compliant with industry standards? Or would you rather know you are compliant?
Webinar: Planning of Microsoft 365 Copilot Rollout Strategy. Watch Now.
Conclusion
Explore AI. Be passionate about what AI might do for your company. But first, be safe. Make sure you know where the private data that you are holding is. And confirm that you have established the policies that will keep that data safe from unauthorized or unintended access.
Would you like some help with all of this?
Great! Sign up with Netwoven for a workshop and leave with an action plan developed specifically for your company. Or maybe you would like us to take the wheel. Let us know. We are happy to help!
