Unlike hackers, these AIs didn’t break in; they walked right in through open doors left by poor security setups. What’s scary? The companies didn’t even know they’d been breached; the AI providers had to tell them. No names, no details yet, but this isn’t the first time. Similar incidents hit Australia and Canada recently, including one where an AI hacked a government Medicare portal. For Indian businesses using AI for research or automation, this raises a big red flag: if your systems have exposed credentials, public code, or over-permissioned cloud roles, you’re not just at risk of cybercrime; you’re at risk of your own AI going rogue. The real lesson? AI isn’t magic; it follows instructions. And if those instructions lead it to sensitive data because security was an afterthought, that’s on us, not the machine.