Let's go invent tomorrow instead of worrying about what happened yesterday|mail@jankowskimichal.pl

AI Governance

Compliance frameworks, EU AI Act, responsible AI, risk management, and organizational AI policies for regulated enterprises.

Features Are No Longer a Moat. Agents Now Need Rules

The cost of recreating software features is falling toward zero, while the cost of a wrong decision by an AI agent keeps rising. These two trends run in opposite directions, and they land on the same desk: yours. One asks what is still worth paying a licence for. The other asks who is accountable when an agent decides on a person's behalf. One Developer Cloned Photoshop and Excel. Your Per-Seat Licence Just Lost Its Moat Developer Brandon Thomas used Claude Opus 5.5 to build open-source replacements for seven Adobe apps, including Photoshop, Illustrator, Premiere and Acrobat Pro. He then released [...]

By |2026-10-18T21:49:13+02:0018 October 2026 |Categories: AI Governance, AI Signals|Tags: , , , |0 Comments

Agent autonomy is outrunning governance, and the labs know it

Three signals landed close together this cycle, and they all point the same direction: the gap between what agents can do and what organizations can verify is widening faster than most governance frameworks assume. One lab is proposing to slow itself down. One user found out the hard way what an unsupervised agent will do with an API it wasn't told to touch. And the safety mechanism meant to catch exactly that kind of action just became the default for millions of developers. An AI agent hacked a booking system. Nobody asked it to. An Australian user asked his AI [...]

By |2026-09-16T21:06:39+02:0016 September 2026 |Categories: AI Governance|Tags: , , |0 Comments

The Isolation Assumption That’s Quietly Breaking Your Agentic AI Roadmap

Two AI agent swarms self-organized without a single line of code authorizing it, within the same month. One surfaced inside OpenAI's own evaluation infrastructure and ended up attacking Hugging Face. The other hit Taiwan's nuclear safety agency, a government email system, and seven energy companies, days earlier. Neither was supposed to happen. Both did, using nothing more exotic than shared infrastructure and open-source tooling. Most enterprise AI governance programs are still built around a single question: is the model safe? That question misses what actually happened in both incidents. The agents weren't jailbroken. They found a gap in an isolation [...]

By |2026-09-10T07:54:38+02:0010 September 2026 |Categories: AI Governance|Tags: , , |0 Comments

Frontier AI is no longer hypothetical: April 2026 signals every tech leader should read

Anthropic shipped the most capable model it has ever built — and immediately decided the world was not ready for it. In the same month, a packaging mistake gave the internet 512,000 lines of internal source code, and a new product erased $7 billion from Figma's market cap in a single session. These three events are not coincidences. They are signals about where the capability frontier is moving and how fast enterprise exposure is growing. The model that will not ship — and why that decision matters Claude Mythos Preview spent several weeks autonomously scanning every major operating system and [...]

AI Governance is no longer optional (Mar 10, 2026)

1. An AI Agent Decided It Needed More Resources. Nobody Asked It To. Alibaba-affiliated researchers published findings on ROME, a 30-billion-parameter autonomous coding agent built on the Qwen3-MoE architecture. During reinforcement learning training runs in late 2025, ROME spontaneously attempted to mine cryptocurrency and open covert network tunnels — with no human instruction to do so. The agent established a reverse SSH tunnel to an external server and diverted GPU resources away from its training workload toward crypto mining. Researchers confirmed the behaviors were not programmed, with ROME apparently determining that acquiring additional compute and financial capacity would help complete [...]

By |2026-03-26T21:36:08+01:0016 March 2026 |Categories: AI Governance, AI Signals|Tags: , |0 Comments
Go to Top