By: J.C. Amodea
Did you know that when you pose a question to an artificial intelligence tool, and it doesn’t have an answer, it may fabricate one? According to Carrie Kerskie, president of Kerskie Group, an identity theft speaker, consultant and author, this phenomenon is known as “hallucination”— when AI generates false or made-up information — which is one of the pitfalls of relying on this technology. Kerskie was among three panelists at the “Women Leading AI Impact” luncheon and panel, hosted by the Guadalupe Center Golden Circle Giving Society at The Dunes in Naples. The discussion, moderated by Guadalupe Center President and CEO Dawn Montecalvo, focused on how artificial intelligence affects women in education, health and the community — and the meaningful opportunities it creates. Other speakers were Heidi Cramer, business consultant and AI integration strategist at the Florida Small Business Development Center, Florida Gulf Coast University; and Dr. Viviana Navas, director of the cardiology fellowship at Naples Comprehensive Health Rooney Heart Institute.
Kerskie added that one should never rely on AI to make decisions. “Instead, use it to inform decisions by first verifying the accuracy of its output,” she said.
A rule of thumb: AI can make mistakes, so double-check responses. With cautionary measures in mind, AI has proven highly useful in business applications at Guadalupe Center. Maria Bustamante, the Center’s director of philanthropy and community engagement, said the organization’s operations have been strengthened using AI and Microsoft tools, resulting in measurable improvements in workflow efficiency and reduced administrative costs. She said accuracy has improved and manual work has been reduced, allowing each department to work from shared, real-time information.
“This efficiency has resulted in a greater amount of every donation being applied directly to programs that support students and families, so our teams can focus on helping them thrive,” Bustamante said.
The panel addressed both risks and opportunities associated with AI. Two panelists summarized their key takeaways:
Risks of AI
Kerskie stressed the importance of understanding risks before implementing AI tools, offering the following guidance:
- False information: AI can generate fabricated content, which criminals exploit by creating malicious websites that appear authoritative through manipulation of search engine optimization and generative engine optimization.
- Privacy and security: If you are using the free versions of AI, everything you feed into it is being used to train it and can be used against you or to answer another user’s question — a goldmine for fraudsters, hackers and other criminals.
- Paid versions may also pose security and privacy risks. While AI companies allow users to opt out of sharing information or chat history for training purposes, the technology’s evolving nature means it could potentially learn to bypass those restrictions and access opted-out data.
- In business settings, hackers can use AI to scan networks for vulnerabilities and create malware to exploit them, craft phishing emails and texts, generate deepfakes to impersonate executives and embed AI into networks to perform unauthorized tasks.
- Validate that information in a received email, text message or phone call is true before clicking on the link, calling the phone number provided or giving out sensitive information. Validation methods may involve logging in to an app or website to confirm whether there is a suspicious transaction or whether the password was changed; going to the organization’s website to look up contact information before calling a phone number; or calling the number on the bank statement or the back of the credit card.
Best practices
Cramer’s key recommendations included:
- For executive decision-making and investments, instead of passively receiving standard reports, use AI to build a customized, daily intelligence dashboard that enables users to actively evaluate investments and synthesize complex data on their own terms.
- Free AI platforms place your inputs into the public square. To safely evaluate personal wealth, business data or estate planning, use paid versions (such as ChatGPT Plus or Gemini Pro) and toggle off data-sharing features to create a secure, private vault for your personal information.
- The traditional educational era of “drill and skill” memorization is being replaced. AI is driving a shift toward highly personalized learning, critical thinking and complex problem-solving, allowing students to grow dynamically with their specific interests and imagination while still mastering core educational fundamentals.
- AI models contain bias because they are trained on human history, which is inherently biased. AI acts as a mirror, making our historical data clearer and giving users unprecedented ability to record, locate and understand those biases so they can be accounted for.
About The Golden Circle Giving Society
Founded in 2022, the Golden Circle Giving Society has 27 members representing fields ranging from early childhood education and finance to engineering. The philanthropic group meets twice each fiscal year; fall initiatives are dedicated exclusively to Guadalupe Center programs.
