Could rushes be key to disproving 'Deepfake' video?

Editorial Type: Opinion Date: 08-2020 Views: 963 Tags: Storage, Compliance, Workflow, Archival, Data protection, Object Matrix PDF Version:
Nick Pearce-Tomenius of Object Matrix looks at some of the potential compliance issues surrounding long term storage of raw footage for TV and media production companies

A recent article in the Guardian raised the possibility of footage on the Jeremy Kyle show having been altered in order to tell the story that the producers wanted to be told, saying: "The family has concerns that the footage is polished and edited, and does not represent the totality of the footage that would have been recorded on all cameras on the day."

The lack of retention of 'rushes' in a drama is unlikely to have a negative impact on society in future years but as the Kyle story highlights the retention of original footage needs to be taken more seriously where factual content is being edited or manipulated.

Another example where studio footage was key in a criminal prosecution is the "Who Wants To Be A Millionaire" cheating case, as Wikipedia recounts: "In court, Ingram claimed the videotape of his appearance on Millionaire was 'unrepresentative of what I heard', and he continues to assert that it was 'unfairly manipulated'. A video recording, with coughing amplified relative to other sounds including Ingram's and Tarrant's voices, was prepared by Celador's editors for the prosecution and 'for the benefit of the jury' during the trial."

Given its nature live action content is difficult to manipulate even with the 'broadcast delay" but not so if the delay is in the minutes, hours, days or months, as is typical for reality based programming. This raises three questions for those producing factual content and also presents a real challenge for those organisations in terms of retaining the potential hundreds of hours of raw footage that goes into producing an hour of finished content:

1. How are production companies and broadcasters protecting rushes or footage captured by studio cameras on the day?
2. Can they prove authenticity of those rushes in the years to come?
3. Is it even possible to retain the original footage and find the clips you need when required?

Protecting rushes/dailies is not new in highly regulated industries like financial institutions that typically are required to adhere to internal or external regulations. They typically have to implement platforms and processes that ensure content security, access control and availability of historical data.

Imagine the scenario where an analyst from a global bank gives an interview where the advice imparted during broadcast differs from the advice given on camera at the time of shooting - advice that might bankrupt individuals, companies or even countries. This manipulation of the message or story can be achieved with subtle editing or more recently the advances in Deepfake technology.

A flip side to Deepfake video or manipulation in the edit is that people, politicians in particular, could use the fact that the technology exists to vehemently deny ever having said or done something on camera, as highlighted in a recent article by Daniel Thomas (BBC News): "The first risk is that people are already using the fact Deepfakes exist to discredit genuine video evidence. Even though there's footage of you doing or saying something you can say it was a Deepfake and it's very hard to prove otherwise."

It would appear that being able to prove the authenticity of raw footage has never been more important.

Production companies who own the IP and rights for shows like "Jeremy Kyle" and "Millionaire" typically rent the studios and pay for the services of post-production companies to get the show made. Those studio and post companies will generally be responsible for protecting the rushes until the show has aired and many will hold on to them for longer periods of time until they no longer have the physical space or resources to manage them.

These cases highlight the need to find content from a show aired several years ago - a task that cannot always be done quickly, if at all. The Jeremy Kyle rushes were protected by the post-production company involved, but that is not always the case. Most organisations simply do not have the technology platforms nor the processes in place.

One of the main concerns will always be "What is the business model?" Keeping finished content in an archive requires resources and long term investment but there is a value in exploiting that content. Doing the same for thousands of hours of raw footage has a less obvious return on investment.

The only way companies will feel compelled to archive rushes forever is via regulation or as an insurance requirement to assist should any future litigation occur. If such regulations are introduced companies will be expected to find and produce evidential content within reasonable time frames or get fined.

Good Digital Content Governance, a mix of process and technology, can ensure that content is protected, instantly accessible and proven to be authentic at any time in the future. It can also help organisations to beat Deepfake or disprove manipulated images.

  • Ensuring content is authentic: DCG platforms make multiple copies of content on ingest using checksums (digital fingerprints) to ensure its integrity from day one and throughout the lifetime of the content. DCG can place retention policies on the data such that not even administrators can accidentally delete it.
  • Protecting data: Digital Preservation processes ensure your content is protected at ingest and remains protected throughout its lifetime. However, this requires regular integrity checking which can be a costly exercise with legacy technology such as LTO. DCG platforms handle all aspects of good digital preservation practice, from continuous content protection and multiple copy protection (on and off-site) to business rules support.
  • Access: Providing searchable audits of every action during the lifetime of the media is essential, as it means you can track exactly what has happened to that content and who has accessed it. DCG platforms offer native, searchable audits of every action from ingest, moves, deletions, attempted deletions and most importantly, read. It has to be said that audit is also possible with public cloud accounts if the user logins are granular to individuals performing the actions.
  • Search: Find is key. With the increasing volume of data in and out of a facility, metadata management is as important as protecting the content itself. The ability to search for content based on up-to-date and relevant metadata will unlock the value of content for many organisations. Loosely coupled metadata and content will always make Find an inefficient or impossible process. DCG platforms protect the metadata along with the essence for the lifetime of the content. Using APIs enables future proof, integrated and automated workflows that ensure content can be found even if media asset management is not available. DCG platforms can also automate the extraction and indexing of any embedded metadata which can vastly increase search efficiency.
  • Business Continuity: Using incumbent platforms that rely on legacy archive and backup practices does not guarantee continuity of business operations. It is a fact that loss of data or access to data can lead to catastrophic loss of revenue for any sized company. DCG platforms provide automated and integrated business continuity functionality ensuring work can continue despite any outages. Implementing automated, asynchronous replication of metadata, data and user access information ensures that everything that is needed will be available at the DR location. Integration of DCG platforms into the end user ecosystem (i.e. they do not have to learn new skills) also makes this a non-disruptive process.

As detailed above, implementing a good DCG platform that is integrated into media workflows will bring value to the organisation and ensure content can be found under any circumstances.

In summary there are some technical, commercial and cultural issues to address in the creative video community if raw footage and archive content is to be protected in accordance with internal or external regulations. One of the biggest challenges will be the physical resources needed to archive thousands of hours of potentially 4 and 8k raw footage.

One potential option is to create a mezzanine or proxy version of those rushes, in a certified transformation workflow, that take up much less space than the originals but retain enough quality for video processing to be applied at future dates. Metadata can be captured during the ingest and transformation process or that metadata can be generated later on using AI platforms.

Keeping those rushes on LTO or SAN/NAS platforms will not be sufficient in terms of good Digital Content Governance nor the ability to efficiently process the files in automated workflows. These rushes will need to be kept in an object storage or cloud storage platform whose automated technologies ensure that good DCG is followed and ensures that rushes are instantly available and searchable.

More info: