There is an interesting debate happening around AI security. Experts seem to have differing opinions on the risks of open-source AI tools being misused.

Some security professionals strongly warn about the potential dangers of OSS. They point out that malicious users could illegally utilize these tools, and that’s where the danger of AI lies. However, here’s the interesting part: when looking at actual data, the story appears somewhat different.

Many researchers highlight that, in reality, most of the AI risks are tied to proprietary systems from major companies like OpenAI and Claude. In other words, the problem isn’t solely about open-source. Furthermore, biosecurity experts have also entered the discussion, arguing that software and sequencing technologies are not the true limiting factors.

In summary, focusing only on open-source when discussing AI risks might be a one-sided view. It’s important to take a more measured look at where the real threats actually lie.