Workplace toxicity is widely recognized as detrimental to organizational culture, yet quantifying its direct impact on operational efficiency remains methodologically challenging due to the ethical and practical difficulties of reproducing conflict in human subjects. This study leverages Large Language Model (LLM) based Multi-Agent Systems to simulate 1-on-1 adversarial debates, creating a controlled "sociological sandbox". We employ a Monte Carlo method to simulate hundrets of discussions, measuring the convergence time (defined as the number of arguments required to reach a conclusion) between a baseline control group and treatment groups involving agents with "toxic" system prompts. Our results demonstrate a statistically significant increase of approximately 25\% in the duration of conversations involving toxic participants. We propose that this "latency of toxicity" serves as a proxy for financial damage in corporate and academic settings. Furthermore, we demonstrate that agent-based modeling provides a reproducible, ethical alternative to human-subject research for measuring the mechanics of social friction.

Online Social Networks (OSNs) widely adopt content moderation to mitigate the spread of abusive and toxic discourse. Nonetheless, the real effectiveness of moderation interventions remains unclear due to the high cost of data collection and limited experimental control. The latest developments in Natural Language Processing pave the way for a new evaluation approach. Large Language Models (LLMs) can be successfully leveraged to enhance Agent-Based Modeling and simulate human-like social behavior with unprecedented degree of believability. Y et, existing tools do not support simulation-based evaluation of moderation strategies. We fill this gap by designing a LLM-powered simulator of OSN conversations enabling a parallel, counterfactual simulation where toxic behavior is influenced by moderation interventions, keeping all else equal. We conduct extensive experiments, unveiling the psychological realism of OSN agents, the emergence of social contagion phenomena and the superior effectiveness of personalized moderation strategies.

Online platforms take proactive measures to detect and address undesirable behavior, aiming to focus these resource-intensive efforts where such behavior is most prevalent. This article considers the problem of efficient sampling for toxicity detection in competitive online video games. To make optimal monitoring decisions, video game service operators need estimates of the likelihood of toxic behavior. If no model is available for these predictions, one must be estimated in real time. To close this gap, we propose a contextual bandit algorithm that makes monitoring decisions based on a small set of variables that, according to domain expertise, are associated with toxic behavior. This algorithm balances exploration and exploitation to optimize long-term outcomes and is deliberately designed for easy deployment in production. Using data from the popular first-person action game Call of Duty: Modern Warfare III, we show that our algorithm consistently outperforms baseline algorithms that rely solely on players' past behavior. This finding has substantive implications for the nature of toxicity. It also illustrates how domain expertise can be harnessed to help video game service operators identify and mitigate toxicity, ultimately fostering a safer and more enjoyable gaming experience.

The rapid advancement of Large Vision-Language Models (LVLMs) has enhanced capabilities offering potential applications from content creation to productivity enhancement. Despite their innovative potential, LVLMs exhibit vulnerabilities, especially in generating potentially toxic or unsafe responses. Malicious actors can exploit these vulnerabilities to propagate toxic content in an automated (or semi-) manner, leveraging the susceptibility of LVLMs to deception via strategically crafted prompts without fine-tuning or compute-intensive procedures. Despite the red-teaming efforts and inherent potential risks associated with the LVLMs, exploring vulnerabilities of LVLMs remains nascent and yet to be fully addressed in a systematic manner. This study systematically examines the vulnerabilities of open-source LVLMs, including LLaVA, InstructBLIP, Fuyu, and Qwen, using adversarial prompt strategies that simulate real-world social manipulation tactics informed by social theories. Our findings show that (i) toxicity and insulting are the most prevalent behaviors, with the mean rates of 16.13% and 9.75%, respectively; (ii) Qwen-VL-Chat, LLaVA-v1.6-Vicuna-7b, and InstructBLIP-Vicuna-7b are the most vulnerable models, exhibiting toxic response rates of 21.50%, 18.30% and 17.90%, and insulting responses of 13.40%, 11.70% and 10.10%, respectively; (iii) prompting strategies incorporating dark humor and multimodal toxic prompt completion significantly elevated these vulnerabilities. Despite being fine-tuned for safety, these models still generate content with varying degrees of toxicity when prompted with adversarial inputs, highlighting the urgent need for enhanced safety mechanisms and robust guardrails in LVLM development.

Large Language Models (LLMs) have demonstrated great capabilities in natural language understanding and generation, largely attributed to the intricate alignment process using human feedback. While alignment has become an essential training component that leverages data collected from user queries, it inadvertently opens up an avenue for a new type of user-guided poisoning attacks. In this paper, we present a novel exploration into the latent vulnerabilities of the training pipeline in recent LLMs, revealing a subtle yet effective poisoning attack via user-supplied prompts to penetrate alignment training protections. Our attack, even without explicit knowledge about the target LLMs in the black-box setting, subtly alters the reward feedback mechanism to degrade model performance associated with a particular keyword, all while remaining inconspicuous. We propose two mechanisms for crafting malicious prompts: (1) the selection-based mechanism aims at eliciting toxic responses that paradoxically score high rewards, and (2) the generation-based mechanism utilizes optimizable prefixes to control the model output. By injecting 1\% of these specially crafted prompts into the data, through malicious users, we demonstrate a toxicity score up to two times higher when a specific trigger word is used. We uncover a critical vulnerability, emphasizing that irrespective of the reward model, rewards applied, or base language model employed, if training harnesses user-generated prompts, a covert compromise of the LLMs is not only feasible but potentially inevitable.

Common problems in playing online mobile and computer games were related to toxic behavior and abusive communication among players. Based on different reports and studies, the study also discusses the impact of online hate speech and toxicity on players' in-game performance and overall well-being. This study investigates the capability of pre-trained language models to classify or detect trash talk or toxic in-game messages The study employs and evaluates the performance of pre-trained BERT and GPT language models in detecting toxicity within in-game chats. Using publicly available APIs, in-game chat data from DOTA 2 game matches were collected, processed, reviewed, and labeled as non-toxic, mild (toxicity), and toxic. The study was able to collect around two thousand in-game chats to train and test BERT (Base-uncased), BERT (Large-uncased), and GPT-3 models. Based on the three models' state-of-the-art performance, this study concludes pre-trained language models' promising potential for addressing online hate speech and in-game insulting trash talk.

Chatbots are used in many applications, e.g., automated agents, smart home assistants, interactive characters in online games, etc. Therefore, it is crucial to ensure they do not behave in undesired manners, providing offensive or toxic responses to users. This is not a trivial task as state-of-the-art chatbot models are trained on large, public datasets openly collected from the Internet. This paper presents a first-of-its-kind, large-scale measurement of toxicity in chatbots. We show that publicly available chatbots are prone to providing toxic responses when fed toxic queries. Even more worryingly, some non-toxic queries can trigger toxic responses too. We then set out to design and experiment with an attack, ToxicBuddy, which relies on fine-tuning GPT-2 to generate non-toxic queries that make chatbots respond in a toxic manner. Our extensive experimental evaluation demonstrates that our attack is effective against public chatbot models and outperforms manually-crafted malicious queries proposed by previous work. We also evaluate three defense mechanisms against ToxicBuddy, showing that they either reduce the attack performance at the cost of affecting the chatbot's utility or are only effective at mitigating a portion of the attack. This highlights the need for more research from the computer security and online safety communities to ensure that chatbot models do not hurt their users. Overall, we are confident that ToxicBuddy can be used as an auditing tool and that our work will pave the way toward designing more effective defenses for chatbot safety.

Modulate, a leader in the fight against online toxicity, announced the completion of a $30 million Series A funding round led by Lakestar with participation from existing investors Everblue Management, Hyperplane Ventures, and others. In addition, Mika Salmi, Managing Partner of Lakestar, will join Modulate's Board of Directors. The company will use the funds to expand its team and continue scaling its groundbreaking proactive voice moderation platform, ToxMod. "ToxMod is changing the way game developers attack toxic behavior in their communities and this funding is a real validation of our mission to make online communities safer," said Mike Pappas, CEO of Modulate. "We're thrilled to welcome Mika and his vast store of experience to the Board as we grow our team and ramp up the development and deployment of ToxMod."

With the US presidential election now 40 days away, all eyes are focused on how online conversations, in conjunction with other hallmarks of online life like viral videos, news clips, and misleading ads, will be used, and often abused, to influence people's decisions. But political discourse, of course, is just one of the ways that user-generated content on the internet is misused for toxic ends. Today, a startup that's using AI to try to tackle them all is announcing some funding. Spectrum Labs -- which has built algorithms and a set of APIs that can be used to moderate, track, flag and ultimately stop harassment, hate speech, radicalization, and some 40 other profiles of toxic behavior, in English as well as multiple other languages -- has raised $10 million in a Series A round of funding, capital that the company plans to use to continue expanding its platform. The funding is being led by Greycroft, with Wing Venture Capital, Ridge Ventures, Global Founders Capital, and Super{set} also participating.

Sam Haberern, 20, was playing Call of Duty on Xbox at his family's house in Connecticut, and he was on a roll. After several dozen high-scoring rounds, other gamers started to take notice. He began receiving invites from players asking him to play with them. He accepted one and joined in the group's online conversation through his headset. "It was great," said Haberern in an interview with The Washington Post.