arXiv subjects

Ilan Strauss

Publications and source records attributed to Ilan Strauss.

4 recordsLinked to original sources

arXiv2025

"Rich-Get-Richer"? Platform Attention and Earnings Inequality using Patreon Earnings Data

Using monthly Patreon earnings, we quantify how platform attention algorithms shape earnings concentration across creator economies. Patreon is a tool for creators to monetize additional content from loyal subscribers but offers little native distribution, so its earnings proxy well for the attention creators capture on external platforms (Instagram, Twitch, YouTube, Twitter/X, Facebook, and ``Patreon-only''). Fitting power-law tails to test for a highly unequal earnings distribution, we have three key findings. First, across years and platforms the earnings tail and distribution exhibits a Pareto exponent around $\alpha \approx 2$, closer to concentrated capital income than to labor income and consistent with a compounding, ``rich-get-richer'' dynamic (Barabasi and Albert 1999). Second, when algorithms tilt more attention toward the top, the gains are drawn disproportionately from the creator ``middle class''. Third, over time, creator inequality across social media platforms converge toward similarly heavy-tailed (and increasingly concentrated) distributions, plausibly as algorithmic recommendations rises in importance relative to user-filtered content via the social graph. While our Patreon-sourced data represents a small subset of total creator earnings on these platforms, it provides unique insight into the cross-platform algorithmic effects on earnings concentration.

econ.GN↗

arXiv2025

The Attribution Crisis in LLM Search Results

Web-enabled LLMs frequently answer queries without crediting the web pages they consume, creating an "attribution gap" - the difference between relevant URLs read and those actually cited. Drawing on approximately 14,000 real-world LMArena conversation logs with search-enabled LLM systems, we document three exploitation patterns: 1) No Search: 34% of Google Gemini and 24% of OpenAI GPT-4o responses are generated without explicitly fetching any online content; 2) No citation: Gemini provides no clickable citation source in 92% of answers; 3) High-volume, low-credit: Perplexity's Sonar visits approximately 10 relevant pages per query but cites only three to four. A negative binomial hurdle model shows that the average query answered by Gemini or Sonar leaves about 3 relevant websites uncited, whereas GPT-4o's tiny uncited gap is best explained by its selective log disclosures rather than by better attribution. Citation efficiency - extra citations provided per additional relevant web page visited - varies widely across models, from 0.19 to 0.45 on identical queries, underscoring that retrieval design, not technical limits, shapes ecosystem impact. We recommend a transparent LLM search architecture based on standardized telemetry and full disclosure of search traces and citation logs.

cs.DL↗

arXiv2025

Real-World Gaps in AI Governance Research

Drawing on 1,178 safety and reliability papers from 9,439 generative AI papers (January 2020 - March 2025), we compare research outputs of leading AI companies (Anthropic, Google DeepMind, Meta, Microsoft, and OpenAI) and AI universities (CMU, MIT, NYU, Stanford, UC Berkeley, and University of Washington). We find that corporate AI research increasingly concentrates on pre-deployment areas -- model alignment and testing & evaluation -- while attention to deployment-stage issues such as model bias has waned. Significant research gaps exist in high-risk deployment domains, including healthcare, finance, misinformation, persuasive and addictive features, hallucinations, and copyright. Without improved observability into deployed AI, growing corporate concentration could deepen knowledge deficits. We recommend expanding external researcher access to deployment data and systematic observability of in-market AI behaviors.

cs.AI↗

arXiv2025

Beyond Public Access in LLM Pre-Training Data

Using a legally obtained dataset of 34 copyrighted O'Reilly Media books, we apply the DE-COP membership inference attack method to investigate whether OpenAI's large language models show recognition of copyrighted content. Our results based on this small sample suggest that GPT-4o, OpenAI's more recent and capable model, exhibits patterns consistent with recognition of pay-walled book content, with an AUROC score of 0.82 (95% bootstrapped CI: 0.60-0.96), though this wide confidence interval reflects substantial uncertainty due to the limited number of books tested. GPT-4o Mini, as a much smaller model, shows little recognition of any O'Reilly Media content with an AUROC score of 0.56 (0.28-0.83) for non-public data. Testing multiple models, with the same cutoff date, provides a partial control for potential language shifts over time that might bias our findings, though differences in model size, architecture, and potentially training data composition limit the strength of this control. These preliminary results underscore the importance of increased corporate transparency regarding pre-training data sources and the development of formal licensing frameworks for AI content training. Our principal contribution is our examination of public and non public data separately.

cs.CL↗