Top AI Players Failing the Safety Test?

The Center for AI Safety's semi-annual AI Safety Index just dropped, and the results are a massive wake-up call. The data paints a sobering picture of where the industry actually stands.

It turns out that even the giants leading the charge in innovation are lagging significantly when it comes to security and existential risk management. Anthropic might be at the top of the list, but even they only managed a C+. Meanwhile, heavyweights like OpenAI and Google DeepMind are sitting at a C, and Meta is dragging behind with a D+.

What’s even more striking is the "F" grades handed out to xAI, DeepSeek, and Mistral. It seems like in the absolute sprint to achieve AGI, everyone is cutting corners on safety frameworks, governance, and accountability.

In my opinion, this highlights the classic "innovation vs. safety" dilemma. Companies are so focused on scaling parameters and compute power that they're treating safety as a secondary feature rather than a core requirement. We are essentially building high-speed engines without checking if the brakes actually work.

If the industry continues to prioritize speed over stability, we might face much more than just technical bugs—we could be looking at systemic risks that are hard to reverse.

What do you guys think? Are these low grades just a result of overly strict metrics, or are we actually heading toward a safety crisis?