• Resources

Secure Transcription: How Safe is Your Data?

By its very nature, a secure transcription often deals with sensitive matters and includes personal, highly confidential or even top-secret information. How do transcription providers ensure that the people handling your data can be trusted to keep it secure? Secure transcriptions are a key procedural tool for formal investigations, from coroners’ inquests and police interviews to regulatory inquiries and employment-related tribunals and hearings. The information being discussed may be personal, commercially sensitive, or critical to the outcome of a criminal trial or a high-stakes fraud investigation, for example. Some of these proceedings have implications at the very highest level. Clearly, this data needs to be handled securely; the consequences of it being widely shared or falling into the wrong hands could be very serious. Most leading transcription providers have rigorous, secure information handling processes in place, to ensure leaks don’t happen, with some accredited to ISO27001 or signed up to the UK government’s Cyber Essentials scheme. But processes are only half the story. What about the people? When it comes to Appen’s official written records, it’s humans rather than machines that turn speech to text. Recording technicians, transcribers, note-takers, and translators all play their part in secure transcription, along with support staff, quality assurance specialists, and managers. In the UK, there are a number of different levels of security clearance designed specifically for those working with sensitive or classified material. These include:
  • Disclosure and Barring Service (DBS) – a check on criminal convictions for people in certain professions.
  • Baseline Personnel Security Standard (BPSS) – a basic pre-employment check aimed primarily at civil servants, the armed forces and government contractors.
  • Security Check (SC) – the most widely held level of vetting, including checks on employment and financial history, spent and unspent criminal convictions and sometimes an interview.
  • Counter-Terrorism Check (CTC) – an enhanced level of vetting for people working with data of value to terrorists or with access to public figures at risk from terrorism.
  • Official Secrets Act (OSA) – usually incorporated into a contract of employment, this offers legal protection against espionage or the unauthorised disclosure of information.
All of these are used in the secure transcription industry, with the level of vetting applied dependent on a person’s role and the type of data they’re dealing with. The best transcription providers carry out basic checks themselves – including verifying identity, home address and employment status – on employees as well as any freelancers working on their behalf. Beyond this, the process is usually client-led, as Caroline Williams, Director at Appen’s Secure Facility in Exeter explains: “We ask clients to identify the level of security they need for each project and then make sure we comply. Typically, clients undertake any higher-level vetting themselves and the clearance usually only applies to that specific job or project.” Providers can be proactive, too. At Appen, everyone working at the company’s secure on-site facility is vetted at least to DBS level, including transcribers, administrative staff, and managers, providing security and confidentiality for your data. In addition, all those working at Appen’s head office in Exeter, as well as associated freelancers, have sections of the OSA included in their contracts. So much formal, official record transcription work is related to the criminal justice system; even when it’s not, the stakes are still high. Data security is paramount and leaks can have profound consequences. Establishing the right data handling systems and processes is important, but ensuring the right people are in place is critical to data security too.   For more information on Appen’s information security, including vetting, contact: enquiries.exeter@appen.com
Website for deploying AI with world class training data
Language