The University of South Carolina's AI Institute hosted the Safe AI for Seniors 2026 workshop Friday. This is the workshop's second year hosting educational sessions designed to help adults ages 55 and older better understand artificial intelligence.
The organizer of the workshop, USC AI Institute and CSE Professor Biplav Srivastava, says he wanted to make AI education accessible to people of all ages, not just younger generations.
“There aren't many avenues for people in their senior years. [Because] of physical and cognitive abilities or inabilities in senior years, we need extra care," he said. “It’s not enough just to release new software and new apps all the time and let people figure it out."
This year's workshop had three, panel-led sessions focusing on cybersecurity, mental health, and data infrastructure.
Experts discussed how seniors can prevent identity theft, phishing and other types of scams. Tips included avoiding disseminating personal information like your social security number or bank information, avoiding links and attachments from suspicious emails, and setting up a network firewall.
Panelists also warned of what they call robocalls. Robocalls use automated dialing systems rather than live callers, or connect recipients to no one at all. Experts say, if this happens to someone, to not say hello because scammers can clone your voice to then use for fraudulent purposes.
The overlap of security and mental health were also discussed. Experts say the use of large language models (LLM) as an emotional companion or to get general information, mental health, or general health advice have become more common in recent years. Professor and Chair of Neuropsychiatry and Behavioral Sciences Meera Narasimhan say research indicates that AI can make people dependent on them emotionally.
"It's about how to maintain a healthy relationship, and that's what, sometime, AI cannot tell us," Narasimhan said. "So, be mindful when using AI, be mindful of the phenomenon that we call sycophancy, meaning that AI sometimes overly validates what we feel."
Srivastava says seniors often experience isolation which can create space for these issues to occur.
"The way that AI technologies are being released to people, it is exacerbating the situation,” Srivastava says. She says this experience of isolation crosses generations, with younger people and seniors exhibiting this vulnerability.
Harsh Vardhan, a middle school student at TLC at Dent Middle School, presented to workshop attendees, offering a youth perspective on AI. He says, "AI is certainly a tool, but [AI] is a tool and a weapon. It's like a kitchen knife. You can use it to cut vegetables, but you can also cut yourself."
Vardhan says it's best to approach AI with levelled curiosity, "you can ask AI [quick questions] or simple tips, but other people can use it to make scams and deep fakes that end up harming you."
Joseph F. Rice School of Law Professor Zoe Niesel warns of the legal implications of sharing personal information to LLM's.
"If you're using a free or general purpose model, know that one reason it's free is because your data is accessible to it," she says.
Niesel recommends using a general purpose AI model in temporary or incognito mode to avoid the model collecting your data to train its models.
"We have a very strong legal structure around health care information in the United States under HIPAA and its [affiliated] statutes. But, the reality is ChatGPT, Claude, Brock, [are] not bound by those kind of things," Niesel said.
As AI continues to evolve, experts say education will be key to helping people navigate both its benefits and its risks.
