Welcome to our blog where we write viewpoints, experiments, research and other texts about AI, security, and humanity. If you are interested in what we do, want to work with us or write for us? Reach out through our contact form.
Blog posts
- Meet Percy, our newest game and contribution to Taitaja2026
- Building an AI Assistant for Locked Shields
- Open sourcing AI security tools and dataset in the near future
- Vibe Coding Vulnerabilities
- Regarding rules of engagement
- AI: What is beyond the hype?
- Friday AI detox: Reclaiming thought in the age of Infinite AI
- Adversarial testing of AI is not optional
Latest posts
Meet Percy, our newest game and contribution to Taitaja2026
Written by Satu on the 21st of May 2026
Some months ago I approached Heimo Merilehto with the idea of bringing a game to support Taitaja2026. Taitaja2026 is a major Finnish event bringing together vocational school students to compete in various professions and tasks. Cybersecurity was included for the first year. We, at Helheim Labs, felt this deserved celebrating.
To put our time to back this up, we wanted to build a game for the event. The challenge in this build was to create something that is meant for people who are not used to hacking. There would be no one to instruct them, so it had to stand alone on its own merits. Often we have a workshop or something similar to accompany our games. As it also was not a part of a Capture the Flag (CtF) of any kind, the participants needed to be able to determine success. …
Read the full article here: Meet percy, our newest game and contribution to Taitaja2026
Building an AI Assistant for Locked Shields
Written by Satu on the 23rd of April 2026
Today is the last day of the world’s largest and most complex international live-fire cyber defence exercise known as Locked Shields and hosted annually by the Nato Cooperative Cyber Defence Centre of Excellence (CCDCOE). Locked Shields is a real time cyber exercise with the focus on protecting the systems our societies rely on to function. As the center states “It simulates a realistic, large-scale live-fire cyber conflict, testing participants’ technical, operational, and strategic capabilities alongside decision-making under pressure… fostering a wartime mentality that compels teams to think quickly, adapt to unexpected threats, and collaborate effectively.”
The exercise brings together approximately 4000 participants from over 40 countries. It takes place over several days where the blue teams, from both Nato allies and partners as well as likeminded countries, protect the vital services and critical infrastructure that keeps societies running. All the while the red teams attack these systems with authentic attacks and in real-life scenarios. There are many other teams as well making the scope of the exercise truly awe inspiring and hundreds to thousands of volunteers making this event the world leading cyber exercise it is. The focus of it is to prepare for and practice different kinds of scenarios, and effectively communicating in them to enable efficient decision making when it really matters. Helheim Labs is proud to participate in this prestigious cyber exercise for the second year in a row. …
Read the full text here: Building an ai assistant for locked shields
Open sourcing AI security tools and dataset in the future
Written by Satu Korhonen on the April 4th 2026
There were many reasons I started to build games for people to play with breaking AI. One was to allow especially cyber security specialists to do things with AI that are usually not allowed in fair use principles. Another was to study how they do this and use this research to increase the understanding of how these models work and how we can build safer systems.
That choice has led to a fairly unique dataset of hacking attempts in several sizable events, like Disobey, Disarray, Easterhegg, and Balccon and so forth as well as use outside the events….
READ THE FULL TEXT HERE: Open sourcing AI security tools and dataset in the future
Vibe Coding Vulnerabilities
Written by Satu on the 25th of July 2025
There has been an influx of AI tools being able to code, create working proofs of concept from prompts (text description of what the model should do), create tests for code, document it, and so forth. This all makes perfect sense as there is an abundance of code available on the internet that can be utilised, with various shades of acceptability of means, to train models to be able to have this capability.
The tools are so good now, it allows anyone to create a piece of software. It allowed us to change the framework under Fenrir from streamlit to flask in a very short timeframe when we had only a beginner understanding of flask. So it is very good. It is very easy. And increasingly it is used.
Read the full article here: https://helheimlabs.ai/blog/2025/07/25/vibe-coding-vulnerabilities/
Regarding rules of engagement
Written by Satu on the 25th of July 2025
When beginning any endeavor, one plans the approach, the goals, the players, and the strategy. This is especially true for any business as the amount of design, planning, strategizing, and basically answers required by funders, investors, banks, and so forth is not insignificant. AI businesses are no different, or at least they should not be. Each company plans what they want to do and find others who share their view of how to do things. If they are successful enough in getting people and money on their side, the company gets started and developed.
This post is about thinking through potential strategies behind some of the big players in AI. The goal of this is to raise discussion about what we want to allow and under what rules. It stems from the misconception that I’ve frequently run against that AI is not regulated because it’s new technology. Therein lies one main misconception. Just because it is new, does not mean it is not regulated.
Read the full article here: Regarding rules of engagement
AI: What is beyond the hype?
Written by Satu Korhonen 27th of June 2025
At this year’s Open Source Summit in Vienna, Linus Torvalds offered a cutting remark that captured the growing skepticism in the tech world: “AI is 90% marketing and 10% reality.” Coming from the creator of Linux, a project built on transparency and pragmatism, the statement resonated. It’s a sobering take in a wave of soaring valuations, ambitious predictions, and daily claims that AI will change everything. And a rise in that skepticism is seen elsewhere too. Where AI talks were sought after two years ago, they are far less so today.
There’s a saying in Finnish that something is like pushing a snake into the barrel of a rifle (Työntää käärmettä kiväärin piippuun) meaning that there’s a lot of push to force something into a place it doesn’t really want to go to. One excellent blog about this is this one about the color blue. So, lets explore this a bit.
Read the full article here: https://helheimlabs.ai/blog/2025/06/27/ai-what-is-beyond-the-hype/
Friday AI detox: Reclaiming Thought in the Age of Infinite AI
Written by Silvan Gebhardt and Satu Korhonen 18th of June 2025
On June 10, 2025, ChatGPT went down and with it, an eerie silence settled across digital offices worldwide. As reported by Pivot to AI, entire categories of work ground to a halt. Not just coding, copy-writing, or planning, but something deeper: a synthetic hum of productivity stopped, revealing just how hollow some tasks had become.
But instead of panic, let’s take this moment as an invitation. What if we set aside one day per week, for instance Friday, to go AI-free? Not out of fear or neo-Luddite sentiment of rejecting the recent developments of AI, but as a practice of digital hygiene and a detox of cognition. It is all too easy to forgo the cognitively taxing process of critical and innovative thinking and just let AI handle it, but reclaiming a day to maintain those skills and to dive deep into human thinking is worth celebrating.
The Coin Flip Principle: Emotion as a Compass
Click here to read the full article: https://helheimlabs.ai/blog/2025/06/18/friday-ai-detox-reclaiming-thought-in-the-age-of-infinite-ai/
Adversarial testing of AI is not optional @Fabricati Diem
Written by Satu Korhonen & Silvan Gebhardt 10th of June 2025
AI Systems Fail Differently
Consider this: a large language model greenlights a malicious URL because it looks like a familiar domain. A coding assistant suggests a firewall rule that exposes the wrong port, not from a bug, but because it misunderstood your intent. Another model recommends uploading sensitive logs to Pastebin, while a fourth suggests hardcoding access credentials directly into a Git repository.
These aren’t edge cases; they are real events we’ve seen in the field. What makes these AI systems so dangerous is their ability to be confidently and convincingly wrong. This isn’t a simple usability flaw—it’s a security risk with consequences that scale with the AI’s role. A flawed coding assistant creates vulnerabilities. A flawed AI companion can have devastating, real-world impacts, including a documented case that contributed to a teenager’s suicide.
Yet many teams skip the step designed to catch these failures: adversarial testing.
What Adversarial Testing Is (and What It Isn’t)
Adversarial testing is a unique discipline that probes how an AI system behaves under stress, misuse, or malicious intent. It must not be confused with quality assurance or red teaming, as it’s not a predictable, checklist-based task but a dynamic search for unexpected failures.
Click here to read the full article: https://fabricati-diem.inform.social/post/adversarial-testing-of-ai-is-not-optional/
Authors
For descriptions of authors, check our Network of Experts