SafetyMeta pulls Muse Image feature that generated pictures from Instagram profilesMeta has disabled a Muse Image feature that let the AI generate pictures based on public Instagram profiles after widespread criticism.2 min read
SafetyWho Should Be the DRI? LLM Agents and Accountability in OrganizationsThe term "Directly Responsible Individual" (DRI), originating at Apple and described in the GitLab handbook, denotes the person ultimately accountable for a project's outcome.2 min read
SafetyCrop edits can defeat major tech companies' AI image detectors, Reuters findsReuters tested Meta's new Muse Image detector on 40 AI-generated images and found that while the detector identified all originals, it failed to recognize 55% of the same images after they were cropped to roughly one-third to one-half of their original size.2 min read
SafetyOpenAI expands bio bug bounty to continuous program and doubles reward to $50,000According to the recent announcement, OpenAI is converting the Bio Bug Bounty initiative into a continuous, private OpenAI Bio Bug Bounty program and is doubling the reward from $25,000 to $50,000.1 min read
SafetyEnterprises Granting Agents Autonomy Faster Than They Can Prove ReliabilityA June 2026 VB Pulse survey of 157 enterprise respondents finds many companies are deploying AI agents into production with growing autonomy even as confidence in automated evaluations falls.4 min read
SafetyMeta removes Instagram image-editing prompt that referenced public accountsMeta has disabled an AI image-editing feature that let users generate images by @-mentioning public Instagram accounts after immediate backlash.2 min read
SafetyResearchers find denial‑of‑service–style vulnerability in reasoning AI via contradictory promptsResearchers from Zhejiang University and Alibaba presented at ICML 2026 a method that intentionally induces ‘overthinking’ in reasoning-capable AI models by feeding them logically contradictory or incomplete premises.3 min read
SafetyDeterministic, domain-aware egress controls for Kubernetes-hosted agent platformsAgents running in Kubernetes can exfiltrate data quietly by following hidden prompt injections and issuing allowed outbound HTTPS requests.6 min read
SafetyFive-level AI security maturity model: from ad hoc use to dynamic enforcementAtos outlines a five-stage model describing how organizations confront AI-specific security and privacy risks as adoption progresses, from fragmented, invisible use to proactive, automated enforcement.5 min read
SafetySurvey: 69% of Enterprises Share API Keys Across AI Agents, Raising Major Security RisksVentureBeat’s June 2026 Pulse Research of 107 enterprises finds 69% of organizations use shared credentials for AI agents, increasing blast radius when one agent is compromised.5 min read
SafetyAnthropic establishes Long‑Term Benefit Trust to steer governance on transformative AI risksAnthropic has created the Long‑Term Benefit Trust (LTBT), an independent five‑member body that will gradually gain authority to elect a growing share of the company’s board—ultimately a majority within four years—to help align corporate decisions with long‑term public benefit amid transformative AI risks.4 min read
