
A San Francisco transcription company used by police departments, prosecutors, government agencies, and Silicon Valley tech firms exposed sensitive customer data online after leaving an internal storage bucket publicly accessible on the web. The lapse at Tigerfish revealed transcriptions of police interviews and body-cam footage, along with raw audio and video files and customer invoices, before the company locked the data down.
Tigerfish data exposure surfaced in an openly accessible cloud bucket
The security issue involved an Amazon-hosted S3 cloud server that was listed publicly and could be reached through an easy-to-guess web address, according to security researcher Anurag Sen, who discovered the files in October. The exposed material included customer transcriptions and the original source materials used to create them, making the lapse especially sensitive given Tigerfish’s client base.
Among the files were transcriptions of police interviews, body-camera footage, and customers’ raw audio and video. The exposed records also included invoices belonging to Tigerfish customers, which included police departments and prosecutors’ offices. That mix of operational documents and source media created a broad privacy and security risk beyond the transcribed text itself.
Sen said no password was required to access the bucket. He later provided details to TechCrunch after he said he did not hear back from the company when he first alerted it to the issue.
Who found it and why the disclosure matters
Sen is a good faith security researcher who has spent years searching the internet for exposed data with the aim of notifying organizations before the information is misused. He describes his work as an “unpaid air-traffic controller,” a metaphor for the nonstop effort involved in identifying sensitive systems that have been accidentally left exposed.
From his island home in Indonesia, Sen has previously uncovered a wide range of sensitive material online, including U.S. military emails exposed from a compartmentalized Department of Defense cloud server, databases with customer records, systems tied to a national license plate surveillance program, and public servers that inadvertently sent out SMS codes used for account logins around the world. His track record is one reason the Tigerfish exposure drew scrutiny: researchers like him often identify problems before criminals do.
After TechCrunch verified the source of the leak and the data, the publication contacted Tigerfish. Soon after, the company secured the exposed S3 bucket.
How Tigerfish responded
Tigerfish president Adam Goldberg said in an email that the company had closed the exposure once it was informed. “We’ve blocked all public access to the bucket you identified, including directory listings and individual files,” Goldberg said. “We’ve also verified that all of our other S3 buckets are blocked to public access.”
Goldberg added that the company is “continuing to review and strengthen our storage security more broadly.” Tigerfish also said Sen’s disclosure email had been caught in its spam filter, which may have delayed the company’s awareness of the issue.
It remains unclear whether Tigerfish has notified its customers about the exposure. The company did not respond to a follow-up email asking about that point.
Why this kind of leak keeps happening
The Tigerfish incident is another example of how a simple cloud misconfiguration can expose highly sensitive data online. Security researchers and cloud providers have repeatedly warned that public storage buckets, missing access controls, and overlooked test environments can leave confidential records open to anyone who knows where to look.
Sen said many of the incidents he finds are not the result of sophisticated attacks, but of basic mistakes. In his view, the internet’s security model still depends too heavily on trust that systems have been configured correctly behind the scenes.
He pointed to common examples such as an engineer spinning up a test database on a Friday, disabling authentication to check an API, and then forgetting to turn access back on before the weekend. Those small errors can have large consequences when the server is directly reachable from the public internet.
AI, vibe-coding, and the old problem of open access
The disclosure also lands at a time when the industry is paying more attention to how AI is changing software development and security. Sen said the rise of vibe-coding and AI-generated development can introduce fresh risks when code is deployed quickly without enough review. But he argued that the most common problems are still old-fashioned human mistakes.
That warning echoes a broader shift among cloud and software companies. Amazon has recently taken steps intended to make it harder for users to make S3 buckets public without passing through several warning screens. Newer vibe-coding platforms such as Lovable and Supabase have also introduced changes designed to reduce user error and prevent accidental exposure.
Even so, the Tigerfish case shows that operational shortcuts and missed configuration checks remain dangerous. When the data involved includes law enforcement interviews, body-cam footage, and customer invoices, the stakes are higher than an ordinary misdirected file share. Sensitive data exposure can affect ongoing investigations, compromise privacy, and erode trust in the organizations that rely on outside vendors.
What the Tigerfish exposure shows
The incident is a reminder that data-handling vendors sit on a large amount of sensitive material and can become a weak point if access controls fail. For companies serving police departments, prosecutors, and major technology firms, storage hygiene is not a back-office detail; it is part of the service itself.
Several practical lessons stand out from the exposure:
- Public cloud storage should be continuously checked for unintended access.
- Security alerts must reach the right inboxes, since spam filters can delay urgent notices.
- Source files, not just processed transcriptions, can contain highly sensitive information.
- Organizations should review all cloud buckets, not just the one that was reported.
- Customer notification processes should be ready in case exposed data includes regulated or sensitive records.
The broader concern is that even without a sophisticated intrusion, sensitive systems can still end up “wide open to the public internet without a password,” as Sen put it. The Tigerfish exposure fits that pattern closely: a publicly accessible storage bucket, a large amount of internal data, and a company that ultimately had to close the door after the fact.
Source: Original report
Was this helpful?
Explore more: Application Audit & Review More Cybersecurity Tech News
Last Modified: November 20, 2026 at 10:32 pm
0 views
