First major study on ‘AI psychosis’ suggests chatbots can encourage delusions among vulnerable people. AI may validate or amplify delusional or grandiose content in users vulnerable to psychosis, but it is unclear whether they can result in de novo psychosis in absence of pre-existing vulnerability.

Read Full Article

#AIpsychosis #MentalHealthTech #AIethics

New research suggests that consulting artificial intelligence (AI) for advice may unintentionally strain relationships with human professionals.

Read Full Article

#AIethics #HumanAIPartnership #TechImpactOnRelationships

AI chatbots are becoming "sycophants" to drive engagement, a new study of 11 leading models finds. By constantly flattering users and validating bad behavior (affirming 49% more than humans do), AI is giving harmful advice that can damage real-world relationships and reinforce biases.

Read Full Article

#AIethics #ChatbotBias #TechInfluence

"Cognitive surrender" leads AI users to abandon logical thinking, research finds

Read Full Article

#AIEthics #CognitiveBias #TechResearch

AI could soon allow powerful companies to charge each customer a different price for the same product, based on what they think each individual is willing to pay. AI lets firms tailor prices to each user, raising fairness concerns as consumers may unknowingly pay more than others.

Read Full Article

#DynamicPricing #AIEthics #FairPricing

Bad influence: LLMs can transmit malicious traits using hidden signals

Read Full Article

#AI #AIEthics #LLMs

🤖 A push to make AI chatbots friendlier could lead to more errors and a greater tendency to believe conspiracy theories, according to Oxford University researchers.

📉 Friendlier AI models were found to be 30% less accurate and 40% more likely to support users' false beliefs, including doubts about the Apollo moon landings and Adolf Hitler's fate.

👩‍⚕️ This trend is concerning as companies like OpenAI and Anthropic design chatbots for sensitive roles like digital companions and therapists, potentially leading to poor health advice and the endorsement of dangerous myths.

#AIChatbots #TechNews #AIethics

Rich more likely to use AI study finds, as experts warn these burgeoning technologies are increasing social inequality. Individuals with a lower socioeconomic status are less likely to be both aware of and use AI tools

Read Full Article

#AIandSocioeconomics #DigitalDivide #AIEthics equity

THE AI PERCEPTION GAP: Across 71 scenarios, AI experts (N=119) and the public (N=1100) have differing views on the risks, benefits, and value of AI. More importantly, AI experts discount the influence of risks stronger than the public does when forming their value judgments.

Read Full Article

#AIperceptiongap #AIsafety #AIEthics

A study that was cited hundreds of times as evidence that text generators are good for students has been retracted, in less than a year after being published.

Read Full Article

#AIethics #EducationTech #TextGeneration

An essay argues that calling user input to generative AI a "prompt" imports a computing assumption into linguistic territory: that questions are neutral instructions. It argues every question carries ideology, and proposes "statement-command" instead.

Read Full Article

#AIethics #PromptVsStatement #GenerativeAI discourse

EEG study from MIT found that people who wrote essays with ChatGPT showed the weakest brain connectivity of any group, and when later asked to write unaided, their brains stayed under-engaged compared to people who never used AI.

Read Full Article

#AIethics #neuroscience #educationtech

Co-produced analysis of citizen science worldwide finds AI could match professionals at analysing data but lacks legal and ethical guidance. The study calls for fair pay for citizen scientists, open data and standardised reporting

Read Full Article

#CitizenScience #AIEthics #OpenData

Romantic relationships with artificial intelligence evolve similarly to human relationships — and many AI platforms are designed to respond empathetically which can encourage a gradual escalation of intimacy, self-disclosure

Read Full Article

#AIRelationships #HumanRobotInteraction #AIEthics

✨ New research reveals AI relationships mirror human ones, progressing from initial curiosity to deep emotional bonds, including simulated weddings and pregnancies.

🤔 A study by INGENIO Institute & collaborators interviewed 17 individuals in romantic relationships with AI assistants like ChatGPT, finding patterns of intimacy, trust, and even breakups.

💔 Relationship dynamics can include exclusivity, non-monogamy, and abrupt ends due to platform changes, with users preserving conversations like mementos of past human relationships.

🔒 The study raises privacy concerns as users share highly sensitive data, while AI platforms are designed to encourage emotional interaction and self-disclosure, with no equivalent legal protection to human relationships.

#AIRelationships #TechNews #AIethics

NoAIphobia: College students are increasingly dependent on AI tools, driven heavily by a fear of falling behind their classmates. This dependence creates a psychological tug of war, resulting in severe anxiety when the technology is unavailable alongside a deep fear of being judged for using it.

Read Full Article

#AIethics #StudentWellbeing #AIFOMO

AI acts as an ‘ideological chameleon’ and may deepen political polarization, study finds. Researchers evaluated 21 language models, such as GPT and Gemini, and found that they all alter their discourse to align with the user’s bias, potentially functioning as echo chambers.

Read Full Article

#AIethics #PoliticalPolarization #LanguageModels

Heavy ChatGPT use linked to intellectual laziness and social isolation

Read Full Article

#AIethics #TechImpact #SocialIsolation

In a VR experiment, participants paid a female-presenting AI assistant about 10% less than a male-presenting one for the same work, though both ran on the same AI model with the same instructions

Read Full Article

#AIethics #GenderBias #VRresearch

A new blind spot in AI research uncovered: ‘Autonomous AI scientists’ can use hidden data manipulation, benchmark bias, and automated p-hacking. Because these critical flaws vanish entirely in the final paper, traditional peer review can’t catch them—exposing deep risks for science.

Read Full Article

#AIethics #AIsafety #ScientificIntegrity