If you enjoy our newsletter, please consider to be a paid subscriber to help us keep more news and updates coming out.
[New] Q2 2026 is available to download here.
Highlights
UN independent science reportSummer 2026: AI Safety ReportAccenture BreachGPT 5.6 launchSingapore Safeguards for Agentic Finance
Deep Dive
Independent Science Report on AI UN
Seek balance risk and opportunity
Rapid development in different domains
Risk and negative impact
Human rights, society system, env
Need evidence for effective governance
Summer 2026: AI Safety Index FutureOfLife
Review top 9 frontier AIs
7 areas of assessments
Top 3: Anthropic, OpenAI, Google
Existential Safety is the worst
Overall: From C+ to F
Accenture Breach BleepingComputer
35GB of data
In July
Source code, company data
Key: RSA, SSH, PAT, access and config
Threat actor “888”
GPT 5.6 launch OpenAI
Strongest Cybersecurity
ExploitBench: 73.5% token budget
ExploitGym: 24.9% peak pass rate
SEC-Brench Pro: 71.2% improved latency
Support cyber defend tasks
SAFR: Safeguards for Agentic Finance at Runtime MAS
Practical industry reference approach
Runtime governance of agentic AI
Define agent authority
Assess actions before execution
Start point for discussion
![[Available] Book Report Q1 2026](https://substackcdn.com/image/fetch/$s_!UrNn!,w_140,h_140,c_fill,f_webp,q_auto:good,fl_progressive:steep,g_auto/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe822165f-0454-4166-8b0b-75582950bf85_1464x834.png)
![[Available] Book Report Q3, 2025](https://substackcdn.com/image/fetch/$s_!HI5v!,w_140,h_140,c_fill,f_webp,q_auto:good,fl_progressive:steep,g_auto/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F62b0655a-9a73-4382-8201-d9007269e7ad_900x900.jpeg)
