AI transcription and data privacy in newsrooms under scrutiny

AI transcription and data privacy in newsrooms under scrutiny

A new report warns that AI transcription tools may be putting newsroom data at risk. As journalists weigh convenience against confidentiality, experts urge the use of secure, enterprise-grade solutions designed to protect sources and uphold public trust.


A report from the Public Media Alliance (PMA) and the University of Edinburgh has raised fresh concerns about the growing use of AI transcription tools in newsrooms, warning that convenience may be coming at the cost of data privacy.

The research found that journalists — particularly those in public service media — are increasingly relying on free or low-cost transcription platforms without full awareness of the security risks involved. Respondents cited fears over data misuse, editorial independence, and the protection of confidential sources.

One case study referenced in the report was the Australian Broadcasting Corporation (ABC), which banned a major AI transcription vendor after it failed key data security and privacy checks.

Although the study focused on public media, experts warn the implications extend across the industry. Newsrooms of all sizes now use AI-driven tools to streamline workflows, raising questions about how sensitive material is stored, shared, and potentially exposed to third parties.

The report underscores the value of adopting enterprise-grade transcription tools that prioritise data protection. These include platforms that do not train on user content, maintain strong encryption standards, and meet regulatory requirements such as the EU’s General Data Protection Regulation (GDPR).

“Security is definitely super important,” said Marion Solletty, Editor-at-Large for France at POLITICO. “We’re talking to confidential sources, some of them very highly placed. We don’t want that to be something that can break out in the open. So source protection is super important for us. When I was trained on Trint, I definitely took note of that.”

Trint, founded by former BBC correspondent Jeff Kofman, provides transcription and translation in over 50 languages, with encryption using TLS 1.2+ and AES 256-bit standards. It offers data residency options in both the EU and US, aligning with GDPR requirements and supporting automatic backup via Amazon Web Services.

For journalists, the report concludes, ensuring source confidentiality and maintaining public trust must take precedence over convenience. As AI becomes embedded in editorial workflows, newsroom leaders face increasing pressure to vet their tools with the same rigour they apply to their reporting.



  • ExtraHop unveils new agentic SOC capabilities

    ExtraHop unveils new agentic SOC capabilities

    ExtraHop launches new tools to power agentic security operations. The network detection and response company has introduced enhanced visibility, identity integrations, and Kubernetes telemetry capabilities designed to provide AI agents within security operations centres with the contextual data required to operate autonomously and respond to advanced threats at machine speed.


  • US court blocks expanded merger disclosure rule

    Federal judge halts expanded US merger disclosure requirements. A Texas court has blocked a rule that would have significantly broadened the information companies must provide in US merger filings, marking a setback for federal antitrust regulators and creating fresh uncertainty around the future direction of disclosure reform.


  • KKR-Arctos deal transcends private equity growth

    KKR-Arctos deal transcends private equity growth

    KKR acquires Arctos Partners for $1.4bn, reshaping private equity. The acquisition offers immediate access to a diversified sports portfolio and highlights a shift in private equity towards more sophisticated structures and governance amid increasing regulatory scrutiny.