Designers Must Help Build AI Safeguards

Patricia emphasizes monitoring AI post-launch isn't enough and urges UX teams to craft safeguards for user safety.

Play episode from 07:45

chevron_right

Transcript

chevron_right

Transcript

Episode notes

In this Halloween special of Future of UX, we explore the disturbing side of digital design where clever UX crosses into manipulation, AI tools launch without proper oversight, and personalization quietly erodes user autonomy.

From real-life cases involving Snapchat’s rogue chatbot to Amazon’s subscription “traps” and a tragic AI-assisted suicide, this episode dives into the UX decisions and AI deployments that turned dark and what designers need to learn from them.

These themes were at the core of powerful conversations in the most recent AI for Designers cohort, which just wrapped last week. If you’re interested in digging deeper into ethical design, human-centered AI, and how to design with integrity — the next cohort starts in January 2026.

Whether you're a UX designer, strategist, or just curious about the future of ethical tech: this is the episode you shouldn’t skip.

In this episode, we cover:

Why dark patterns still dominate — and why they may soon become illegal
What happens when AI tools roll out too fast (with no safety net)
The emotional toll of designing features you don’t believe in
How personalization is creating echo chambers and limiting human curiosity
Why “seamless” UX might be the most dangerous design trend of all
Actionable advice for spotting (and stopping) unethical product decisions

🔗 Sources & Real-World Cases Referenced in the Episode:

FTC sues Amazon for dark patterns in Prime sign-up and cancelation (FTC, June 2023)
Hilton agrees to disclose hidden resort fees after state AG probe (Nebraska Attorney General, Jan 2024)
Snapchat’s My AI chatbot gave inappropriate advice to a 13-year-old (Washington Post, March 2023)
UnitedHealth sued for using AI with 90% error rate to deny care (Reuters, Nov 2023)
Belgian man dies by suicide after conversations with AI chatbot (Euronews, March 2023)
TikTok’s algorithm creates extremist echo chambers in hours (The Guardian, July 2024)
Meta’s Galactica AI shut down after 3 days due to misinformation (VentureBeat, Nov 2022)
33 U.S. states sue Meta over youth mental health harms (Reuters, Oct 2023)
EU Digital Services Act bans dark patterns across digital platforms (Schönherr EU Law, Feb 2025)
California CPRA makes dark patterns illegal in data privacy contexts (AdExchanger, Aug 2022)