Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
129 tokens/sec
GPT-4o
28 tokens/sec
Gemini 2.5 Pro Pro
42 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Accounting for AI and Users Shaping One Another: The Role of Mathematical Models (2404.12366v1)

Published 18 Apr 2024 in cs.LG, cs.CY, cs.GT, and cs.IR

Abstract: As AI systems enter into a growing number of societal domains, these systems increasingly shape and are shaped by user preferences, opinions, and behaviors. However, the design of AI systems rarely accounts for how AI and users shape one another. In this position paper, we argue for the development of formal interaction models which mathematically specify how AI and users shape one another. Formal interaction models can be leveraged to (1) specify interactions for implementation, (2) monitor interactions through empirical analysis, (3) anticipate societal impacts via counterfactual analysis, and (4) control societal impacts via interventions. The design space of formal interaction models is vast, and model design requires careful consideration of factors such as style, granularity, mathematical complexity, and measurability. Using content recommender systems as a case study, we critically examine the nascent literature of formal interaction models with respect to these use-cases and design axes. More broadly, we call for the community to leverage formal interaction models when designing, evaluating, or auditing any AI system which interacts with users.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (92)
  1. Roles for computing in social change. In FAT* ’20: Conference on Fairness, Accountability, and Transparency, Barcelona, Spain, January 27-30, 2020, pages 252–260. ACM, 2020.
  2. Online recommendations for agents with discounted adaptive preferences. CoRR, abs/2302.06014, 2023.
  3. The economics of recommender systems: Evidence from a field experiment on movielens. In Proceedings of the 24th ACM Conference on Economics and Computation, EC 2023, London, United Kingdom, July 9-12, 2023, page 117. ACM, 2023.
  4. A game-theoretic approach to recommendation systems with strategic content providers. In Advances in Neural Information Processing Systems (NeurIPS), pages 1118–1128, 2018.
  5. Learning with exposure constraints in recommendation systems. In Ying Ding, Jie Tang, Juan F. Sequeda, Lora Aroyo, Carlos Castillo, and Geert-Jan Houben, editors, Proceedings of the ACM Web Conference 2023, WWW 2023, Austin, TX, USA, 30 April 2023 - 4 May 2023, pages 3456–3466, 2023.
  6. Content provider dynamics and coordination in recommendation ecosystems. In Hugo Larochelle, Marc’Aurelio Ranzato, Raia Hadsell, Maria-Florina Balcan, and Hsuan-Tien Lin, editors, Advances in Neural Information Processing Systems 33: Annual Conference on Neural Information Processing Systems 2020, NeurIPS 2020, December 6-12, 2020, virtual, 2020.
  7. Modeling attrition in recommender systems with departing bandits. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pages 6072–6079, 2022.
  8. Manipulation-proof machine learning. CoRR, abs/2004.03865, 2020.
  9. Modeling recommender ecosystems: Research challenges at the intersection of mechanism design, reinforcement learning and generative models. CoRR, abs/2309.06375, 2023.
  10. Diversified recommendations for agents with adaptive preferences. Advances in Neural Information Processing Systems, 35:26066–26077, 2022.
  11. Bandits meet mechanism design to combat clickbait in online recommendation. CoRR, abs/2311.15647, 2023.
  12. Artificial intelligence, algorithmic recommendations and competition. Social Science Research Network, May 2023.
  13. Estimating and penalizing induced preference shifts in recommender systems. In International Conference on Machine Learning, ICML 2022, 17-23 July 2022, Baltimore, Maryland, USA, volume 162 of Proceedings of Machine Learning Research, pages 2686–2708. PMLR, 2022.
  14. Recommender systems and competition on subscription-based platforms. Social Science Research Network, April 2023.
  15. Stochastic bandits with delay-dependent payoffs. In International Conference on Artificial Intelligence and Statistics, pages 1168–1177. PMLR, 2020.
  16. User strategization and trustworthy algorithms. CoRR, abs/2312.17666, 2023.
  17. How algorithmic confounding in recommendation systems increases homogeneity and decreases utility. In Proceedings of the 12th ACM Conference on Recommender Systems. ACM, 2018.
  18. Alexandra Chouldechova. Fair prediction with disparate impact: A study of bias in recidivism prediction instruments. Big Data, 5(2):153–163, 2017.
  19. US Equal Employment Opportunity Commission et al. The americans with disabilities act and the use of software, algorithms, and artificial intelligence to assess job applicants and employees, 2022.
  20. Feedback effects between similarity and social influence in online communities. In Proceedings of the 14th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Las Vegas, Nevada, USA, August 24-27, 2008, pages 160–168. ACM, 2008.
  21. Towards psychologically-grounded dynamic preference models. In Proceedings of the 16th ACM Conference on Recommender Systems, pages 35–48, 2022.
  22. Fairness is not static: deeper understanding of long term fairness via simulation studies. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, pages 525–534, 2020.
  23. Preference dynamics under personalized recommendations. In EC ’22: The 23rd ACM Conference on Economics and Computation, Boulder, CO, USA, July 11 - 15, 2022, pages 795–816, 2022.
  24. SARDINE: A simulator for automated recommendation in dynamic and interactive environments. CoRR, abs/2311.16586, 2023.
  25. Performative recommendation: Diversifying content via strategic incentives. volume 202 of Proceedings of Machine Learning Research, pages 9082–9103. PMLR, 2023.
  26. User tampering in reinforcement learning recommender systems. CoRR, abs/2109.04083, 2021.
  27. Last switch dependent bandits with monotone payoff functions. In International Conference on Machine Learning, ICML 2023, 23-29 July 2023, Honolulu, Hawaii, USA, volume 202 of Proceedings of Machine Learning Research, pages 10265–10284. PMLR, 2023.
  28. Cirs: Bursting filter bubbles by counterfactual interactive recommender system. ACM Transactions on Information Systems (TOIS), 42(1), 2023. ISSN 1046-8188.
  29. Learning and incentives in user-generated content: multi-armed bandits with endogenous arms. In Robert D. Kleinberg, editor, Innovations in Theoretical Computer Science, ITCS ’13, Berkeley, CA, USA, January 9-12, 2013, pages 233–246. ACM, 2013.
  30. Incentivizing high-quality user-generated content. In Sadagopan Srinivasan, Krithi Ramamritham, Arun Kumar, M. P. Ravindra, Elisa Bertino, and Ravi Kumar, editors, Proceedings of the 20th International Conference on World Wide Web, WWW 2011, Hyderabad, India, March 28 - April 1, 2011, pages 137–146. ACM, 2011.
  31. Tarleton Gillespie. The relevance of algorithms. Media technologies: Essays on communication, materiality, and society, 167(2014):167, 2014.
  32. Strategic classification. In Proceedings of the 2016 ACM Conference on Innovations in Theoretical Computer Science, Cambridge, MA, USA, January 14-16, 2016, pages 111–122. ACM, 2016.
  33. Youtube, the great radicalizer? auditing and mitigating ideological biases in youtube recommendations. CoRR, abs/2203.10666, 2022.
  34. Excessive use of online video streaming services: Impact of recommender system use, psychological factors, and motives. Computers in Human Behavior, 80:220–228, 2018.
  35. Recommending to strategic users. CoRR, abs/2302.06559, 2023.
  36. Recommending and evaluating choices in a virtual community of use. In Proceedings of the SIGCHI conference on Human factors in computing systems, pages 194–201, 1995.
  37. Modeling content creator incentives on algorithm-curated platforms. CoRR, abs/2206.13102, 2022.
  38. Incentivizing high-quality content in online recommender systems. CoRR, abs/2306.07479, 2023.
  39. Collaborative filtering for implicit feedback datasets. In Proceedings of the 8th IEEE International Conference on Data Mining (ICDM 2008), December 15-19, 2008, Pisa, Italy, pages 263–272. IEEE Computer Society, 2008.
  40. Matching of users and creators in two-sided markets with departures. CoRR, abs/2401.00313, 2024.
  41. Recsim: A configurable simulation platform for recommender systems. CoRR, abs/1909.04847, 2019.
  42. Clickbait vs. quality: How engagement-based optimization shapes the content landscape in online platforms. CoRR, abs/2401.0980, 2024.
  43. Alternative microfoundations for strategic classification. In Marina Meila and Tong Zhang, editors, Proceedings of the 38th International Conference on Machine Learning, ICML 2021, 18-24 July 2021, Virtual Event, volume 139 of Proceedings of Machine Learning Research, pages 4687–4697. PMLR, 2021.
  44. Supply-side equilibria in recommender systems. CoRR, abs/2206.13489, 2022.
  45. Improved bayes risk can yield reduced social welfare under competition. CoRR, abs/2306.14670, 2023.
  46. Degenerate feedback loops in recommender systems. In Vincent Conitzer, Gillian K. Hadfield, and Shannon Vallor, editors, Proceedings of the 2019 AAAI/ACM Conference on AI, Ethics, and Society, AIES 2019, Honolulu, HI, USA, January 27-28, 2019, pages 383–390. ACM, 2019.
  47. Preference amplification in recommender systems. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining, pages 805–815, 2021.
  48. Bandits with deterministically evolving states. CoRR, abs/2307.11655, 2023.
  49. How do classifiers induce agents to invest effort strategically? In Anna Karlin, Nicole Immorlica, and Ramesh Johari, editors, Proceedings of the 2019 ACM Conference on Economics and Computation, EC 2019, Phoenix, AZ, USA, June 24-28, 2019, pages 825–844. ACM, 2019.
  50. Inherent trade-offs in the fair determination of risk scores. In Christos H. Papadimitriou, editor, 8th Innovations in Theoretical Computer Science Conference, ITCS 2017, January 9-11, 2017, Berkeley, CA, USA, volume 67, pages 43:1–43:23, 2017.
  51. The challenge of understanding what users want: Inconsistent preferences and engagement optimization. In EC ’22: The 23rd ACM Conference on Economics and Computation, Boulder, CO, USA, July 11 - 15, 2022, page 29. ACM, 2022.
  52. Recharging bandits. In 2018 IEEE 59th Annual Symposium on Foundations of Computer Science (FOCS), pages 309–319. IEEE, 2018.
  53. Grouplens: Applying collaborative filtering to usenet news. Communications of the ACM, 40(3):77–87, 1997.
  54. Do offline metrics predict online performance in recommender systems? CoRR, abs/2011.07931, 2020.
  55. Hidden incentives for auto-induced distributional shift. CoRR, abs/2009.09153, 2020.
  56. A last switch dependent analysis of satiation and seasonality in bandits. In International Conference on Artificial Intelligence and Statistics, pages 971–990. PMLR, 2022.
  57. Fine-tuning games: Bargaining and adaptation for general-purpose models. CoRR, abs/2308.04399, 2023.
  58. Engineering a safer recommender system. In Workshop on Responsible Decision Making in Dynamic Environments, 2022.
  59. Rebounding bandits for modeling satiation effects. Advances in Neural Information Processing Systems, 34:4003–4014, 2021.
  60. A field test of bandit algorithms for recommendations: Understanding the validity of assumptions on human preferences in multi-armed bandits. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems, pages 1–16, 2023.
  61. Rotting bandits. Advances in neural information processing systems, 30, 2017.
  62. Delayed impact of fair machine learning. In Proceedings of the 35th International Conference on Machine Learning, ICML 2018, Stockholmsmässan, Stockholm, Sweden, July 10-15, 2018, volume 80 of Proceedings of Machine Learning Research, pages 3156–3164. PMLR, 2018.
  63. Optimal recommendations under attraction, aversion, and social influence. In Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining, pages 811–820, 2014.
  64. Bayesian incentive-compatible bandit exploration. Operations Resesearch, 68(4):1132–1161, 2020.
  65. Feedback loop and bias amplification in recommender systems. In CIKM ’20: The 29th ACM International Conference on Information and Knowledge Management, Virtual Event, Ireland, October 19-23, 2020, pages 2145–2148. ACM, 2020.
  66. Eric Meyerson. Youtube now: Why we focus on watch time, 2012. URL https://blog.youtube/news-and-events/youtube-now-why-we-focus-on-watch-time/.
  67. From optimizing engagement to measuring value. In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, pages 714–722, 2021.
  68. Optimizing long-term social welfare in recommender systems: A constrained matching approach. In Proceedings of the 37th International Conference on Machine Learning, ICML 2020, 13-18 July 2020, Virtual Event, volume 119 of Proceedings of Machine Learning Research, pages 6987–6998. PMLR, 2020.
  69. Recsim NG: toward principled uncertainty modeling for recommender ecosystems. CoRR, abs/2103.08057, 2021.
  70. Katie Notopoulos. Ai spam is already starting to ruin the internet. Business Insider, 2024. URL https://www.businessinsider.com/ai-spam-google-ruin-internet-search-scams-chatgpt-2024-1.
  71. Dissecting racial bias in an algorithm used to manage the health of populations. Science, 366(6464):447–453, 2019.
  72. A classification of feedback loops and their relation to biases in automated decision-making systems. In Proceedings of the 3rd ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization, EAAMO 2023, Boston, MA, USA, 30 October 2023 - 1 November 2023, pages 7:1–7:14. ACM, 2023.
  73. Feedback loops with language models drive in-context reward hacking. CoRR, abs/2402.06627, 2024.
  74. Kari Paul. The rebel group stopping self-driving cars in their tracks – one cone at a time. The Guardian, 2023.
  75. Performative prediction. In Proceedings of the 37th International Conference on Machine Learning, ICML 2020, 13-18 July 2020, Virtual Event, volume 119 of Proceedings of Machine Learning Research, pages 7599–7609. PMLR, 2020.
  76. Billy Perrigo. The new ai-powered bing is threatening users. that’s no laughing matter. Time Magazine, 2023. URL https://time.com/6256529/bing-openai-chatgpt-danger-alignment/.
  77. Recovering bandits. Advances in Neural Information Processing Systems, 32, 2019.
  78. Models for understanding and quantifying feedback in societal systems. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency, pages 1765–1775, 2022.
  79. Recogym: A reinforcement learning environment for the problem of product recommendation in online advertising. CoRR, abs/1808.00720, 2018.
  80. Nir Rosenfeld. Strategic ML: how to learn with data that ’behaves’. In Proceedings of the 17th ACM International Conference on Web Search and Data Mining, WSDM 2024, Merida, Mexico, March 4-8, 2024, pages 1128–1131. ACM, 2024.
  81. Alvin E. Roth. The Theory and Practice of Market Design. https://www.nobelprize.org/uploads/2018/06/roth-lecture.pdf, 2018.
  82. Planning for autonomous cars that leverage effects on human actions. In Robotics: Science and systems, volume 2, pages 1–9. Ann Arbor, MI, USA, 2016.
  83. Artificial intelligence and bank credit analysis: A review. Cogent Economics & Finance, 10(1):2023262, 2022.
  84. Learning to suggest breaks: sustainable optimization of long-term user engagement. In International Conference on Machine Learning, pages 29671–29696. PMLR, 2023.
  85. Nick Seaver. Computing taste: algorithms and the makers of music recommendation. University of Chicago Press, 2022.
  86. Social information filtering: Algorithms for automating “word of mouth”. In Proceedings of the SIGCHI conference on Human factors in computing systems, pages 210–217, 1995.
  87. Reinforcement learning: An introduction. MIT press, 2018.
  88. Jan C Willems. Models for dynamics. Dynamics reported: a series in dynamical systems and their applications, pages 171–269, 1989.
  89. Mixed negative sampling for learning two-tower neural networks in recommendations. In Companion of The 2020 Web Conference 2020, Taipei, Taiwan, April 20-24, 2020, pages 441–447. ACM / IW3C2, 2020.
  90. How bad is top-k recommendation under competing content creators? In International Conference on Machine Learning, ICML 2023, 23-29 July 2023, Honolulu, Hawaii, USA, volume 202 of Proceedings of Machine Learning Research, pages 39674–39701. PMLR, 2023a.
  91. Rethinking incentives in recommender systems: Are monotone rewards always beneficial? CoRR, abs/2306.07893, 2023b.
  92. Sampling-bias-corrected neural modeling for large corpus item recommendations. In Proceedings of the 13th ACM Conference on Recommender Systems, RecSys 2019, Copenhagen, Denmark, September 16-20, 2019, pages 269–277. ACM, 2019.
Citations (3)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now
Youtube Logo Streamline Icon: https://streamlinehq.com

YouTube