Most company leaders are on guard as private information is shared with Ai tools, but most are unaware Ai can corrupt the information they use to make decisions. Chris Draper, founder of Morrigan Ai says innocent Ai use today, can cement inaccurate information tomorrow in company databases, CRM tools, project management, and other tools forever.
Consider every time an employee uses an Ai notetaker and casually uploads those notes to company tools. If they didn’t verify accuracy first, which most employees don’t do, other tools and employees will later be relying on less-accurate information. Draper explains the risk of shadow Ai and steps you should take right away.
What is Shadow Ai?
Shadow Ai is a real problem too of course. Draper warns that Iowa business leaders are facing a silent crisis: “shadow AI.” With 60% to 90% of employees using unsanctioned tools like ChatGPT or Claude, organizations are inadvertently courting massive data breaches. It’s a concern for sure, but while many worry about privacy, the more insidious threat is the “informational sovereignty” problem.
AI is a “human accelerant and amplifier” that can corrupt the very data used to conduct business by generating synthetic information that looks accurate but contains a “guaranteed wrong” rate of roughly 16%.
The Risk of Synthetic Realities
For small to mid-sized businesses, the danger isn’t just a technical glitch; it is financial ruin due to bad information.
Client projects, purchase orders, customer service concerns, trouble shooting, addresses, employee information… if just 5% of this information was bad that could spell financial ruin for a company. Sadly, Ai’s information failure rate is 15%.
Draper illustrates this with a plumbing company using AI for bidding. If the tool is 15% off on margins at a speed the human can’t catch, the company goes bankrupt faster than ever before, even if the physical work is perfect.
Draper notes that Ai is fundamentally different from traditional tech like video conferencing because “the data itself isn’t anywhere near as important as the refactoring, the recreating, and the reconditioning and the remaking of it into new information.” This “spew speed” creates a comfort level where humans stop rechecking work, allowing errors to become permanent parts of the organizational record.
Shifting From IT to Human Resources
Draper advocates for a strategy where AI is viewed as a human resource issue rather than just a software purchase. He suggests that rigid “boxes” used to define employee roles must be broken down to allow for human augmentation. “I will not ever be able to succeed in the tech department with AI as long as human resources is forcing the same boxes,” Draper explains. Instead of blocking tools, leaders should guide employees toward “human-centric” integration that allows them to shift focus away from mundane processes toward high-value relationships. Ultimately, Draper believes that “our access to business in the state of Iowa comes down to our ability to own the information that comes out of our data.”
Thanks to Dentons Davis Brown for collaborating with The Iowa Business Podcast to bring incredible value to our listeners! Collaborators compensate our hosts directly for their time and effort crafting exclusive content.
