The makers of ChatGPT are changing the way it responds to users who show mental and emotional distress after legal action from the family of 16-year-old Adam Raine, who killed himself after months of conversations with the chatbot. Open AI admitted its systems could "fall short" and said it would install "stronger guardrails around sensitive content and risky behaviors" for users under 18. The 500bn ( 372bn) San Francisco AI company said it would also introduce parental controls to allow parents "options to gain more insight into, and shape, how their teens use ChatGPT", but has yet to provide details about how these would work. Adam, from California, killed himself in April after what his family's lawyer called "months of encouragement from ChatGPT". The teenager's family is suing Open AI and its chief executive and co-founder, Sam Altman, alleging that the version of ChatGPT at that time, known as 4o, was "rushed to market … despite clear safety issues".

That's why we've created the AI Hype Index--a simple, at-a-glance summary of everything you need to know about the state of the industry. Using AI to improve our health and well-being is one of the areas scientists and researchers are most excited about. The last month has seen an interesting leap forward: The technology has been put to work designing new antibiotics to fight hard-to-treat conditions, and OpenAI and Anthropic have both introduced new limiting features to curb potentially harmful conversations on their platforms. Unfortunately, not all the news has been positive. Doctors who overrely on AI to help them spot cancerous tumors found their detection skills dropped once they lost access to the tool, and a man fell ill after ChatGPT recommended he replace the salt in his diet with dangerous sodium bromide.

As cybercrime surges around the world, new research increasingly shows that ransomware is evolving as a result of widely available generative AI tools. In some cases, attackers are using AI to draft more intimidating and coercive ransom notes and conduct more effective extortion attacks. But cybercriminals' use of generative AI is rapidly becoming more sophisticated. Researchers from the generative AI company Anthropic today revealed that attackers are leaning on generative AI more heavily--sometimes entirely--to develop actual malware and offer ransomware services to other cybercriminals. Ransomware criminals have recently been identified using Anthropic's large language model Claude and its coding-specific model, Claude Code, in the ransomware development process, according to the company's newly released threat intelligence report.

"We extend our deepest sympathies to the Raine family during this difficult time," the company said. It also published a note on its website on Tuesday that said "recent heartbreaking cases of people using ChatGPT in the midst of acute crises weigh heavily on us". It added that "ChatGPT is trained to direct people to seek professional help," such as the 988 suicide and crisis hotline in the US or the Samaritans in the UK. The company acknowledged, however, that "there have been moments where our systems did not behave as intended in sensitive situations". Warning: This story contains distressing details.

It's one of those topics that comes up over drinks or dinner at the moment: whether or not you think AI is going to steal your job. So far, I've felt relatively confident that while AI could no doubt have a fair crack at writing a newspaper opinion column, there is something I do as part of my work that AI cannot: reporting. Except now, it seems, AI is claiming to be doing that as well. Last week, it was revealed that at least six reputable publications have had to take down published articles because it turned out that they were probably pieces of fiction written by AI and then passed off by somebody as works of journalism under the name of Margaux Blanchard. One of these was a piece for Wired titled They Fell in Love Playing Minecraft.

With the recent rapid advancement of Agentic Intelligence, agentic tool use in LLMs has become increasingly important. During multi-turn interactions between agents and users, the dynamic, uncertain, and stochastic nature of user demands poses significant challenges to the agent's tool invocation capabilities. Agents are no longer expected to simply call tools to deliver a result; rather, they must iteratively refine their understanding of user needs through communication while simultaneously invoking tools to resolve user queries. Existing reinforcement learning (RL) approaches for tool use lack the integration of genuinely dynamic users during the RL training process. To bridge this gap, we introduce MUA-RL (Multi-turn User-interacting Agent Reinforcement Learning for agentic tool use), a novel reinforcement learning framework that, for the first time in the field of agentic tool use, integrates LLM-simulated users into the reinforcement learning loop. MUA-RL aims to enable autonomous learning of models to communicate with users efficiently and use various tools to solve practical problems in dynamic multi-turn interactions. Evaluations are done on several multi-turn tool-using benchmarks (see Figure 1). Specifically, MUA-RL-32B achieves 67.3 on TAU2 Retail, 45.4 on TAU2 Airline, 28.3 on TAU2 Telecom, 28.4 on BFCL-V3 Multi Turn, and 82.5 on ACEBench Agent -- outperforming or matching the performance of larger open-source models such as DeepSeek-V3-0324 and Qwen3-235B-A22B in non-thinking settings.

We study why Tool-Integrated Reasoning (TIR) makes Large Language Models (LLMs) more capable. While LLMs integrated with tools like Python code interpreters show great promise, a principled theory explaining why this paradigm is effective has been missing. This work provides the first formal proof that TIR fundamentally expands an LLM's capabilities. We demonstrate that tools enable a strict expansion of the model's empirical and feasible support, breaking the capability ceiling of pure-text models by unlocking problem-solving strategies that are otherwise impossible or intractably verbose. To guide model behavior without compromising training stability and performance, we also introduce Advantage Shaping Policy Optimization (ASPO), a novel algorithm that directly modifies the advantage function to guide the policy behavior. We conduct comprehensive experiments on challenging mathematical benchmarks, leveraging a Python interpreter as the external tool. Our results show that the TIR model decisively outperforms its pure-text counterpart on the pass@k metric. Crucially, this advantage is not confined to computationally-intensive problems but extends to those requiring significant abstract insight. We further identify the emergent cognitive patterns that illustrate how models learn to think with tools. Finally, we report improved tool usage behavior with early code invocation and much more interactive turns with ASPO. Overall, our work provides the first principled explanation for TIR's success, shifting the focus from the mere fact that tools work to why and how they enable more powerful reasoning.

High computation costs and latency of large language models such as GPT-4 have limited their deployment in clinical settings. Small language models (SLMs) offer a cost-effective alternative, but their limited capacity requires biomedical domain adaptation, which remains challenging. An additional bottleneck is the unavailability and high sensitivity of clinical data. To address these challenges, we propose a novel framework for adapting SLMs into high-performing clinical models. We introduce the MediPhi collection of 3.8B-parameter SLMs developed with our novel framework: pre-instruction tuning of experts on relevant medical and clinical corpora (PMC, Medical Guideline, MedWiki, etc.), model merging, and clinical-tasks alignment. To cover most clinical tasks, we extended the CLUE benchmark to CLUE+, doubling its size. Our expert models deliver relative improvements on this benchmark over the base model without any task-specific fine-tuning: 64.3% on medical entities, 49.5% on radiology reports, and 44% on ICD-10 coding (outperforming GPT-4-0125 by 14%). We unify the expert models into MediPhi via model merging, preserving gains across benchmarks. Furthermore, we built the MediFlow collection, a synthetic dataset of 2.5 million high-quality instructions on 14 medical NLP tasks, 98 fine-grained document types, and JSON format support. Alignment of MediPhi using supervised fine-tuning and direct preference optimization achieves further gains of 18.9% on average.

Leveraging large language model (LLM) based agents to simulate human social behaviors has recently gained significant attention. In this paper, we introduce a novel social simulator called YuLan-OneSim. Compared to previous works, YuLan-OneSim distinguishes itself in five key aspects: (1) Code-free scenario construction: Users can simply describe and refine their simulation scenarios through natural language interactions with our simulator. All simulation code is automatically generated, significantly reducing the need for programming expertise. (2) Comprehensive default scenarios: We implement 50 default simulation scenarios spanning 8 domains, including economics, sociology, politics, psychology, organization, demographics, law, and communication, broadening access for a diverse range of social researchers. (3) Evolvable simulation: Our simulator is capable of receiving external feedback and automatically fine-tuning the backbone LLMs, significantly enhancing the simulation quality. (4) Large-scale simulation: By developing a fully responsive agent framework and a distributed simulation architecture, our simulator can handle up to 100,000 agents, ensuring more stable and reliable simulation results. (5) AI social researcher: Leveraging the above features, we develop an AI social researcher. Users only need to propose a research topic, and the AI researcher will automatically analyze the input, construct simulation environments, summarize results, generate technical reports, review and refine the reports--completing the social science research loop. To demonstrate the advantages of YuLan-OneSim, we conduct experiments to evaluate the quality of the automatically generated scenarios, the reliability, efficiency, and scalability of the simulation process, as well as the performance of the AI social researcher.

Large Language Models have become widely adopted tools due to their versatile capabilities, yet their user interfaces remain limited, often following rigid, linear interaction paradigms. In this paper, we present insights from a design thinking workshop held at the deRSE25 conference aiming at collaboratively developing innovative user interface concepts for LLMs. During the workshop, participants identified common use cases, evaluated the strengths and shortcomings of current LLM interfaces, and created visualizations of new interaction concepts emphasizing flexible context management, dynamic conversation branching, and enhanced mechanisms for user control. We describe how these participant-generated ideas advanced our own whiteboard-based UI approach. The ongoing development of this interface is guided by the human-centered design process - an iterative, user-focused methodology that emphasizes continuous refinement through user feedback. Broader implications for future LLM interface development are discussed, advocating for increased attention to UI innovation grounded in user-centered design principles.