MacVoices #26002: Talking AI and LLMs with The Long Island Macintosh Users Group (2)

Jan 5, 2026 | Interview

Chuck Joiner, Long Island Macintosh Users Group

Real-world risks and responsible use of AI kick off the second part of our conversation with The Long Island Macintosh Users Group. The group swaps scam stories (spoofed bank calls, W-2 phishing, ransomware) and how AI can amplify fraud. Ways to mitigate exposure in an AI-powered world include cyber insurance, privacy tradeoffs in popular AI tools, copyright/IP guardrails in image generation, and careful experimentation.

This edition of MacVoices is supported by MacVoices After Dark. What happens before and after the shows is uncensored, on-topic, off-topic, and always off the wall. Sign up as a MacVoices Patron and get access!
http://patreon.com/macvoices

Show Notes:

Chapters:

[0:00] Scams refresher: don’t click, verify directly
[2:18] Why scams work: volume, pressure, and “sensibility”
[3:44] Spoofed bank calls and how to break the script
[5:49] Small businesses as targets; cyber insurance gap
[7:12] Photo scanning business: liability vs. cyber coverage
[11:35] W-2 breach fallout; IRS PINs and identity theft
[13:42] Ransomware economics and “references” story
[20:13] LLM choices: ChatGPT vs. Perplexity; citations and accuracy
[21:58] Scraping, paywalls, and plagiarism concerns
[26:16] Privacy tradeoffs and risk assumptions
[31:06] Apple, encryption backdoors, and trust
[34:56] Human review triggers; sensitive prompts
[38:33] Closing: experiment, but stay cautious
[40:50] Synthetic hosts, AI conversations, and credibility risks
[41:10] AI companions and teen harms; responsibility and guardrails
[42:15] Phones in schools; education over bans
[46:51] Wrap-up and thanks