AI tools enable Chinese hackers in a rapid campaign

NBC News highlighted a recent report from CrowdStrike indicating that China‑linked hackers leveraged an AI model, DeepSeek, to launch attacks against more than 200 organizations in just 11 days. The speed and scale of the campaign show how readily AI can be weaponized when the right model is accessible.

For responsible AI providers, this incident reinforces the need for strict controls on model usage. At ELP GPT we never train our proprietary models on customer conversations without explicit notice and consent, and we retain complete visibility over any third‑party inference we incorporate.

The incident also raises the bar for security‑by‑design practices. We require explicit approval before any AI action runs inside a customer’s application, ensuring that the customer remains in control of what the model does and when. This approach mitigates the risk of unintended behavior or misuse.

Ultimately, the lesson is clear: as AI capabilities expand, so must accountability, transparency, and consent. Those that embed these principles into their operating model will better protect both their clients and the broader ecosystem. — Edgar Melo Costa, founder

Founder’s notes are opinion. They are not investment, legal or professional advice.