An Indian AI startup that helps companies construct customized chatbots has leaked nearly 350,000 delicate information after the info was left unsecured on the net.
Ahmedabad-headquartered WotNot left a large assortment of delicate consumer data – together with scans of passport and identification paperwork, medical data, resumes, journey itineraries and extra – unsecured in a misconfigured Google Cloud Storage bucket.
Researchers at Cybernews uncovered the safety downside on August 27 2024. The Google Cloud Storage bucket it discovered was storing 346,381 information – all accessible to anyone on the web, no password required.

That lack of even essentially the most primary safety is woeful when you think about that the knowledge contained within the wide-open storage bucket included paperwork that may make it straightforward for a cybercriminal to commit identification theft.
Cybernews tried to tell WotNot of the issue on September ninth, and despatched “a number of follow-up emails, together with to different e mail addresses ” In accordance with the researchers, it took greater than two months for the enterprise to shut the safety gap.
WotNot informed Cybernews that the bucket was utilized by free-tier customers of its providers, and that “the trigger for the breach was that the cloud storage bucket insurance policies have been modified to accommodate a particular use case. Nevertheless, we regretfully missed completely verifying its accessibility, which inadvertently left the info uncovered.”
The AI chatbot firm tried to reassure its enterprise prospects that they weren’t impacted by the safety breach:
“For enterprise prospects, we offer personal cases to make sure safety and compliance requirements are strictly adhered to.”
Frankly, it should not matter in case you are a non-paying consumer of WotNot or an organisation like Merck or the College of California that the corporate lists amongst its paying prospects. No-one deserves to have their privateness handled so recklessly.
By some means I doubt that WotNot was promoting one of many advantages of being a paid-up consumer, moderately than sticking with the free tier, was that there was no safety in place for many who weren’t paying prospects.
My recommendation? By no means share delicate data with an AI chatbot, as you’ll be able to’t be certain the place it is perhaps saved or what might be executed with it… and within the case of providers like WotNot you could not know the way a lot care it’s taking to maintain it out of the arms of actually anybody else on the web.