Skip to content
At nVision, (our annual technology conference) we gathered IT leaders to discuss two major topics: AI and security. In this blog series, our panelists, each with unique industry backgrounds and strategies, will share their insights on today’s top of mind questions.
Jump to meet our panelists section.
Don’t miss the first blog in our byte sized insights series!
While using the AI tools, what limitations are you/end users, running into? Are those limitations being imposed by you, organically within the organization, or is it a limitation of not understanding the tool set?
Jon:
One of our limitations is how fast it’s changing. Our environment right now is within our own tenant. We have a corporate AI policy that sets guardrails for appropriate use, and who gets a license to use which tool. Internally, we have a small sample set of people using co-pilot, but we can’t really manage personal use of ChatGPT or various AIs out there. What we can manage is within our own organization. The biggest thing is education: making sure our content stays within our own tools.
Jason:
The first thing that we wanted to do was back up usage limitations by policy, so we formed a Digital Innovations Committee, with an AI-specific focus. We wrote a policy and rolled it out to the end users, to make sure we set the guard rails outright. This ensures protection of our intellectual property: our end users are very clear on what data they can and cannot put into a chatbot.
Joel:
Limitations that we have with AI right now are primarily self-imposed. We’re on that threshold of being very reluctant to start using AI anywhere, especially where proprietary or sensitive data is concerned. And it’s going to be awhile before we look to break that pattern. Another other limitation is, in our industry, we’re tied to some more discreet software designed specifically for our industry. So, the adoption of AI within that software, and some of those applications, is just going to be a lot longer road than it would be elsewhere.
Jay:
As others have said, you need to establish guardrails, have it clearly written down, and then have everybody sign off on them. Now we all have a clear understanding of what is allowed and what is not, and we have established ramifications if a policy is violated. We also keep logs of what our users are asking AI to do, so if a policy is violated, we have proof in the log files and we can revoke permissions. We also need to anticipate questions and be ready for a curve ball sometimes. We try to document questions and additional information from users, and then use them to dig down even further to see why they are asking those questions, what do they want to use it for, etc.
Follow along in this series as we dive into more questions from today’s IT leaders. Want the rest of your Q&A insights right away? Watch the 45-minute panel recording on demand now.
Curious about AI for your organization? Start a conversation with our team today.
Interested in attending nVision?
Save your free seat today and join us for our biggest event of the year!
Meet our Panelists:
Joel Weber has been with Vector Windows, of Fergus Falls, for about 12 years. He currently leads technology, cyber security, and physical security operations. Vector Windows has between 100 to 160 end users.
Jason LeZalla has been with Microbiologics, Inc. for 2 ½ years and is the IT operations manager. They have 270 users, and 500 computing devices due to the science equipment and all the instrumentation they have.
Jon Artz is with KLJ Solutions Holding Co, a civil engineering company of about 600, and is spread across 26 different offices.
Jay Tambornino is with the Minnesota Bankers Association (MBA Consulting Group). They have 17 employees in their office and advocate on behalf of banks in Minnesota.
