AI Ethics with Fexingo: Bias, Safety, and Responsible Artificial Intelligence

In this episode of AI Ethics with Fexingo, Lucas and Luna explore a disturbing case: a customer-support AI copilot accidentally sharing a user's private chat history with a stranger. They break down how retrieval-augmented generation (RAG) systems can leak sensitive data when context windows are misconfigured, and what that means for trust in enterprise AI. Drawing on a real incident from early 20...

Episode 53 of AI Ethics with Fexingo examines the privacy risks of AI-powered mental health chatbots. Lucas and Luna discuss a 2025 study where an AI therapist platform inadvertently shared session data with advertisers, citing a case involving a user whose therapy transcripts were used to target ads for antidepressants and sleep aids. They explore how current regulations like HIPAA in the US and...

Episode 52 of AI Ethics with Fexingo dives into how AI grading tools in art education can penalize students for creative expression, using a real case from a university digital arts program where an AI consistently gave low scores to works with unconventional brushstrokes. Lucas and Luna explore how training data biases can suppress artistic diversity, what happened when students appealed, and the...

Episode 51 explores a new frontier of algorithmic bias: predictive policing systems that assign risk scores to neighborhoods based on historical arrest data, creating feedback loops that reinforce over-policing in communities of color. Lucas and Luna break down a 2025 study from the AI Now Institute analyzing Palantir's Gotham platform in three mid-sized cities. They uncover how 'hot spot' predict...

Self-driving cars are already on public roads in several US cities, but who decides what the car does when an accident is unavoidable? In this episode, Lucas and Luna explore the trolley problem as it applies to autonomous vehicles. They examine the MIT Moral Machine study, which logged over 40 million decisions from people worldwide, revealing that different cultures have starkly different prefer...

Episode 49 of AI Ethics with Fexingo explores the controversial use of AI to predict life expectancy for insurance underwriting. Lucas and Luna dissect a 2025 case where a major European insurer used a machine learning model trained on lifestyle data, genetic markers, and social media activity to set premiums. They examine the accuracy claims, the regulatory backlash from the EU's AI Act enforceme...

Episode 48 of AI Ethics with Fexingo dives into a surprisingly common and deeply frustrating bias: accent misidentification in voice AI. Lucas and Luna explore a 2025 Stanford study showing that speech-to-text systems from major providers misrecognize words from speakers with African American Vernacular English and Southern U.S. accents up to 30 percent more often than from standard American Engli...

Lucas and Luna dive into a growing but overlooked AI ethics problem: legal AI tools that give confident but wrong answers. They examine a real case from early 2026 where a small business owner relied on an AI legal document generator—only to find the contract it produced was unenforceable in her state. The hosts explore why these errors happen (training data gaps, lack of jurisdiction awareness),...

In this episode of AI Ethics with Fexingo, hosts Lucas and Luna explore the hidden fees embedded in AI-powered investment advisors. They break down how algorithms from major platforms like Wealthfront and Betterment can steer clients into higher-cost funds or generate unnecessary trades to boost platform revenue, often without transparent disclosure. The conversation centers on a real case from 20...

Episode 45 of AI Ethics with Fexingo drills into a real case from this past April: a property management firm in Cleveland used an AI system to flag lease violations and auto-generate eviction notices. Over 200 tenants received notices without a single human verifying the AI's judgment. We trace how the system worked, what went wrong, and why this matters for the push toward algorithmic tenant scr...

Episode 44 of AI Ethics with Fexingo examines a growing concern in healthcare AI: algorithm developers with financial ties to the treatments their models recommend. Lucas and Luna dissect a 2025 study from the Journal of Medical Ethics that found 30 percent of clinical AI tools are co-developed by pharmaceutical companies whose drugs the algorithms prioritize. They explore the case of an AI sepsis...

In this episode of AI Ethics with Fexingo, Lucas and Luna explore the growing use of AI-powered 'people analytics' tools that rank and predict employee performance based on passive data—email patterns, Slack messages, calendar activity—often without explicit consent. They examine a 2025 case where a Fortune 500 retailer used a system from a vendor called Workforce AI to flag 'low-potential' worker...

Episode 42 of AI Ethics with Fexingo examines the growing use of AI grading assistants in college classrooms. Lucas and Luna break down a specific case from Georgia State University, where an automated essay-scoring system was found to reward keyword stuffing and penalize creative writing. They discuss how the system—trained on past student essays—developed a bias toward formulaic structure and ag...

Lucas and Luna examine the risks of AI-powered symptom checkers and health chatbots that provide medical advice without regulatory oversight. They focus on a 2025 study from Stanford Medicine that tested four popular symptom-checker apps—including WebMD's Symptom Checker and Babylon Health—and found they gave correct diagnoses only 32% of the time, while recommending potentially harmful actions 18...

Episode 40 of AI Ethics with Fexingo: Lucas and Luna explore the legal risks of over-relying on generative AI for professional advice. Using the recent case of a lawyer who submitted fake citations from ChatGPT, they examine why AI models hallucinate legal and medical references, how the fine print disclaims responsibility, and what listeners can do to protect themselves. Specific discussion of th...

Lucas and Luna examine how AI systems used in hiring and performance reviews can learn from past interactions and hold negative associations against candidates. They explore a 2025 study from the MIT Media Lab showing that a simulated recruiter AI penalized candidates whose profiles were similar to previously rejected applicants, even when those candidates had objectively stronger qualifications....

Episode 38 of AI Ethics with Fexingo. Lucas and Luna examine the real-world consequences of automated proctoring software used in remote exams. They focus on a specific case: a university student who was flagged for eye movement by an AI proctor, had her exam invalidated, and spent months appealing with no human oversight. The hosts discuss the accuracy of these systems, bias in gaze detection, an...

Episode 37 of AI Ethics with Fexingo explores a hidden bias in hiring algorithms: using estimated commute distance as a proxy for job stability or socioeconomic status. Lucas and Luna dissect a 2025 Stanford study that found AI recruitment tools penalized applicants from lower-income neighborhoods by flagging longer commutes as a risk factor—even when those candidates were more qualified. They dis...

Episode 36 of AI Ethics digs into the privacy nightmare of smart home devices used for childcare. Lucas and Luna examine a recent incident where a popular smart speaker recorded a family's private conversation and sent it to a random contact. They discuss the lack of transparency in how these devices process audio, the loopholes in current privacy laws, and what parents can do to protect their chi...

Episode 35 of AI Ethics with Fexingo dives into the growing use of artificial intelligence in insurance claims processing. Lucas and Luna examine a specific case: a homeowner whose basement flood claim was automatically denied by an AI model that misread the policy language. The hosts discuss how these systems are trained on historical claims data that may encode racial and socioeconomic biases, l...

Episode 34 of AI Ethics with Fexingo investigates the growing practice of automated resume screening—where AI systems reject job applicants before any human recruiter ever sees their qualifications. Lucas and Luna examine a 2025 study from Harvard Business School that found 72% of large employers now use some form of AI pre-screening, and drill into the case of a Fortune 500 company whose algorith...

Lucas and Luna explore the growing use of fully automated loan underwriting systems that leave borrowers with no human appeal process. They examine a 2025 study from the Consumer Financial Protection Bureau that found 78% of denied loan applicants never learn why their application was rejected by an AI system. The hosts discuss the case of a small business owner in Ohio whose repeat loan applicati...

Episode 32 of AI Ethics with Fexingo examines the unregulated world of AI-powered mental health chatbots. Lucas and Luna discuss a recent study finding that popular therapy AIs exhibit racial and socioeconomic bias in their responses. They dive into the case of a chatbot that recommended different coping strategies based on a user's implied income level, and explore why these systems lack clinical...

Lucas and Luna unpack the growing use of AI in rental housing decisions, focusing on a 2025 class action against a major property management firm accused of using an opaque algorithm to screen tenants and recommend evictions. They walk through how the system worked, the data it used (including social media sentiment scores), and what the lawsuit revealed about lack of transparency and disparate im...

In this episode, Lucas and Luna examine a troubling aspect of generative AI: the use of non-consensual intimate images in training datasets. They focus on a 2025 Stanford Internet Observatory report that found more than 3,000 such images in the LAION-5B dataset, which has been used to train models like Stable Diffusion. The hosts discuss how this happened, why it matters for privacy and consent, a...