Generative Artificial Intelligence (Gen AI) is all the rage these days. Odds are, someone in your organization is already using it. If you’re not sure if it’s being used in your organization, it is, and someone is probably paying for a premium version out of their own pocket and might even be expensing it. It is one of the most common instances of shadow IT in the industry right now. From a practical perspective, why are people using it, why should the leadership team care, and what are some pitfalls?
Why Are People Using It?
The most common use case these days is to generate or parse content. Generative AI is still a very new technology, but it can already do some amazing things.
- Don’t feel like reading that 3-page PDF from a vendor on what their product does? Throw it in ChatGPT and ask it to summarize the content for you.
- Need to create an email for your marketing campaign that explains simply and concisely what your company does and why customers should choose your company to be their preferred vendor? Ask Jasper to whip that up for you.
- Need to take good meeting notes but don’t want extra staff in the meeting or awkward pauses while you get your notes caught up? Upgrade to Teams Premium and turn on the Intelligent meeting recap.
- Need a spreadsheet that lists all your pay periods for your 2026 budgets? Ask Copilot for M365 and you shall receive.
Why Should Leadership Care?
AI can save your organization time and money. That got your attention, I bet! Unfortunately, there is a shadow side to that statement, and they are all laid out in the pitfalls section below. Another reason leadership should care is that most organizations don’t have any formal strategy or policy on the use of AI. Without a corporate policy explaining what is and isn’t allowed, anything is allowed.
The barrier for entry to using AI is not high and there is a good chance your employees have already realized that. AI doesn’t need to have an application installed. Anyone can go to ChatGPT or Copilot or any number of other web sites and start using it via their web browser. Subscriptions start well under $10/month and many people see enough value to pay for it themselves and try to expense it.
What Are Some of The Pitfalls?
AI Hallucinations
You ever run into a kid that just can’t admit when they are wrong or don’t know so they just make something up and hope you don’t notice? That is AI today. That lie AI told is referred to as a hallucination. Languages are hard. Anything highly structured, like a programming language or taking existing content and reducing or expanding on that content, is almost a slam dunk right now. Searching for data that may not exist or extrapolating from incomplete and/or complex information is still very problematic. The more abstract or complex the topic is, the more careful you have to be when relying on the results.
Liability
In the United States, courts have already ruled repeatedly that companies are responsible for what their AI applications say. If you publish something your AI wrote or if you use AI chatbots for customer service, your company can be held liable for that information.
Unintended Information Disclosure
Not every AI platform promises to keep things private and, in fact, the default stance for most AI engines is that any information you put into the AI engine can be used to train the engine. In other words, everything you enter into an AI engine is now effectively public. Most AI platforms have a premium option that can limit the information’s use but not every end user understands that fact, knows where those settings are, knows exactly what information is considered private by the company, and/or is using a premium AI account that has those controls available. Some AI platforms, like Copilot for Microsoft 365, will also access the data in your organization (e.g. email in Exchange Online, documents in SharePoint Online) and use that to supplement answers. If your organization doesn’t have security and privacy controls in place, it is entirely possible that information on sensitive topics (e.g. employee compensation) may become accessible to everyone in the organization.
Interested in simplifying implementation or learning how to leverage AI for your business? |
Want to see more articles like this?
Get great content delivered to your inbox!
