Newsletter
Prof. Nitin Agarwal Speaks with Apprenticely on Engineered Social Media Engagement through AI
Following a landmark court ruling holding Meta and YouTube liable for negligent platform design, COSMOS Research Center founding director Prof. Nitin Agarwal provided expert insight in an interview with Apprenticely to discuss the mechanisms driving social media addiction. He detailed how AI-driven recommendation algorithms exploit psychological triggers and eliminate natural stopping cues to maximize time-on-platform and advertising revenue. Drawing on extensive research supported by federal partners such as the U.S. National Science Foundation and the U.S. Department of War, Prof. Agarwal highlighted how these engagement-focused systems are routinely weaponized by malicious actors to amplify outrage and wage cognitive warfare. He noted that growing policy momentum and regulatory pressure, including the Arkansas Governor’s initiatives to focus on youth online safety, could compel platforms to adopt algorithmic transparency, duty-of-care standards, and healthier design frameworks.
COSMOS Takes Center Stage at 20th AAAI ICWSM CySoc 2026 in Los Angeles!
COSMOS presented three groundbreaking studies at the 20th International AAAI Conference on Web and Social Media (ICWSM), Cyber Social Threats (CySoc 2026) in Los Angeles, USA. As a premier global venue hosted by the Association for the Advancement of Artificial Intelligence (AAAI), ICWSM and CySoc bring together international leaders to tackle critical digital risks, including hate speech, radicalization, and online manipulation. Supported by major federal grants focused on strategic regional stability in the Indo-Pacific, COSMOS’s research addresses complex socio-political movements, cross-lingual threat propagation during political crises like the youth protests in Nepal, and algorithmic recommendation pathways during major socio-political events in Taiwan. By combining epidemiological models, multilingual crisis analysis, and graph-mining frameworks, this body of work delivers scientifically grounded solutions to protect the global information environment.
Hot Off The Press: A Smarter Way to Stop Online Toxicity Campaigns
COSMOS recently published a new study in Springer’s Journal of Social Network Analysis and Mining that offers platforms a far more effective way to curb online toxicity. While traditional content moderation targets individual posts or accounts, coordinated toxic campaigns easily bypass these rules by spreading harmful content through tightly connected groups. This research introduces a mathematical optimization approach using Integer Programming to pinpoint and remove focal toxic structures (the core groups driving the harm). Tested on a massive dataset of over 324,000 Telegram users from the Russia-Ukraine conflict, the new approach proved twice as effective at breaking up toxic networks and removing harmful content while using far fewer moderation resources.
Research Spotlight: Understanding Traps in AI-powered Recommendation Algorithms
We spotlight three studies published at the 14th International Conference on Complex Networks and their Applications (Complex Networks), New York, USA, that examine how recommendation systems can push users into traps, shape pathways of influence, and reinforce patterns of engagement. The work introduces new socio-computational approaches for modeling user attraction, measuring trap intensity, and analyzing how persuasive features within homogeneous content can strengthen algorithmic content cycles. Together, these studies advance interpretable methods for auditing AI-based recommendation algorithms and for understanding how digital platforms influence what users discover, consume, and continue to engage with.
From COSMOS to Meta – Again: Ridwan Builds Production-Scale AI
COSMOS doctoral student and graduate assistant Ridwan Amure is currently completing his second internship at Meta, Inc. (Facebook’s parent company), where he is working as a machine learning engineer on product-focused advertising and recommendation systems. Drawing on his research experience under Prof. Nitin Agarwal at COSMOS, Ridwan is applying machine learning, large-scale data analysis, and model evaluation skills to real-world AI systems. His internship is providing valuable experience in developing and evaluating models for production environments, while further expanding his expertise in building AI systems that are accurate, efficient, scalable, and have real-world applications.
New Hire: Vida Torgbe hired as Graduate Assistant at COSMOS Research Center
COSMOS welcomes Vida Torgbe as a new Graduate Assistant, where she contributes to research on social media behaviors and communication efforts. She is currently pursuing an MS in Information Science at UA Little Rock and brings a multidisciplinary background in social sciences, data analysis, program evaluation, and monitoring and evaluation. At COSMOS, she works with Prof. Nitin Agarwal to strengthen her skills in social media analytics, data science, and interdisciplinary research while contributing to projects that address complex societal challenges.