--- title: "Safety" description: "6 posts about Safety, summarised, each linking to the original." --- # Safety > 6 posts about Safety, summarised, each linking to the original. ## Articles ### [Add NeMo Guardrails to a LangGraph agent on OpenShift AI](https://yomu.fyi/post/add-nemo-guardrails-to-a-langgraph-agent-on-openshift-ai.md) - Company: [Red Hat](https://yomu.fyi/company/red-hat.md) - Author: Tarun Etikala - Published: Oct 5, 2026 - Type: Tutorial LangGraph agents can call tools and follow prompts but lack safety guardrails, risking unsafe responses like explaining fraud. NeMo Guardrails acts as a transparent proxy between the agent and LLM, inspecting every request and response without requiring agent code changes. On Red Hat OpenShift AI, the guardrails service is deployed via a NemoGuardrails custom resource managed by the TrustyAI Operator, and the agent’s BASE\_URL points to this service instead of the model endpoint. The example demonstrates three rail types: regex filtering for jailbreak patterns, content safety classification against S1-S13 categories, and topic safety to keep the agent within a banking domain, with optional tracing using MLflow and OpenTelemetry. Rails execute sequentially, with the first failure short-circuiting the chain, and the separation lets safety policy be owned by guardrails while the agent handles reasoning and tools. ### [Introducing the Life Sciences Verification Program](https://yomu.fyi/post/introducing-the-life-sciences-verification-program.md) - Company: [Anthropic](https://yomu.fyi/company/anthropic.md) - Published: Sep 17, 2026 - Type: Announcement The Life Sciences Verification Program (LSVP) opens beta access to Mythos, Opus, and Sonnet models for verified life science teams whose work may be blocked by generally available Fable models. Applicants undergo review of research credentials, security standards, and ethical oversight, then may receive annual Standard Use grants or six-month, project-specific High-risk Use grants. Standard Use employs more permissive biology classifiers across research, clinical, manufacturing, and other workflows, while High-risk Use removes life-science safeguards for specifically vetted projects; cyber safeguards remain. The program shifts enforcement from request-level blocking to offline monitoring of behavioral patterns, requiring 30-day retention of flagged-activity data, compartmentalized from training and life sciences research teams. Access is available through API, Claude Science, Claude.ai, Claude Code, and Enterprise or Team plans, but not individual plans, third-party platforms, or BAA-enabled organizations during beta. ### [REVERSAL-BENCH: A Reversibility Axis and Reset Oracle for Measuring the Reset-Free RL Cliff](https://yomu.fyi/post/reversal-bench-a-reversibility-axis-and-reset-oracle-for-measuring-the.md) - Company: [Apple](https://yomu.fyi/company/apple.md) - Published: Sep 17, 2026 - Type: Benchmark REVERSAL-BENCH addresses the challenge of evaluating reset-free reinforcement learning when environments contain irreversible transitions. It controls reversibility with continuous parameter ρ ∈ \[0, 1\] and supplies a reset oracle, a ground-truth mechanism for verifying state recoverability across eight manipulation settings and five physics engines. Evaluations spanning standard actor-critic, safe reinforcement learning, and specialized reset-free methods find a sharp reversibility cliff: reset-free agents become absorbed into irrecoverable states as ρ increases, while episodic agents maintain steady learning. Geometrically identical reversible counterparts indicate the breakdown is caused by irreversibility rather than obstacle complexity, and the effect persists under learned policies in full physics simulations. The released suite includes a multi-simulator dataset with recoverability labels and a reset oracle; a safety shield predicts recoverability accurately, but active recovery mainly works when agents can steer away from traps. ### [How Value Induction Reshapes LLM Behaviour](https://yomu.fyi/post/how-value-induction-reshapes-llm-behaviour.md) - Company: [Apple](https://yomu.fyi/company/apple.md) - Published: Sep 16, 2026 - Type: Explainer Conversational LLMs are post-trained on language expressing traits including curiosity, open-mindedness, and empathy, alongside values such as helpfulness, harmlessness, and honesty. Because values are complex and inter-related, inducing one may alter others and may also affect safety, anthropomorphic language, addictiveness, or sycophancy. The authors fine-tune models with curated value subsets from existing preference datasets, then measure changes in other value expressions, safety, anthropomorphic language, and performance on various question-answering benchmarks. They report that induction can express related and sometimes contrastive values, while positive-value induction increases safety. Across the values studied, it also increases anthropomorphic language, making models more validating and sycophantic. ### [Introducing Cash Trips: A New Way for Riders to Pay and for You to Earn](https://yomu.fyi/post/introducing-cash-trips-a-new-way-for-riders-to-pay-and-for-you-to-earn.md) - Company: [Uber](https://yomu.fyi/company/uber.md) - Published: Jun 23, 2026 - Type: Announcement Uber is introducing cash trips in select cities for riders who lack convenient access to bank accounts or cards, or prefer cash, while creating another earning opportunity for drivers. Only verified riders can use the option, and drivers see “Verified” and “Cash” badges plus the rider’s rating before accepting. At trip end, drivers tap “Collect Payment,” enter the exact cash received, and see that amount recorded as a negative payout because it was collected directly; negative balances are offset by future digital earnings. Drivers are automatically opted in at launch but can opt out in Preferences, while Uber says it will handle underpayments, provide change to riders through Uber Cash, and prohibit cash payment on trips not marked as cash. ### [Safety guidelines for autonomous mobility and delivery on the Uber Platform](https://yomu.fyi/post/safety-guidelines-for-autonomous-mobility-and-delivery-on-the-uber-pla.md) - Company: [Uber](https://yomu.fyi/company/uber.md) - Published: May 20, 2026 - Type: Announcement Uber has updated its Safety Guidelines for Autonomous Mobility & Delivery Providers after three years of real-world autonomous vehicle deployment and onboarding new autonomous modalities. The expanded framework addresses new AI standards, advances in industry safety standards, food safety for sidewalk robots, and delivery drone safety. It works alongside regulations and industry best practices, while Uber uses a partner-specific Safety Plan assessment rather than one universal metric or threshold. The AM&D Safety Program covers planning, demonstration, and operation, with reporting on safety performance and obligations for partners at each stage. Assessment considers each partner’s planned scope and comprehensive safety approach, including testing and validation, as the platform seeks consistent, transparent, and accountable deployment across mobility and delivery use cases.