Avid updates Europe’s leading broadcasters on AI development

What if artificial intelligence (AI) enabled you to search for something, without actually searching?

What if, rather than trying to remember some obscure, but cool, feature you once heard about, you could just ask an application to do it?

What if you could do that in your own language?

And what if, by doing this, you could be more creative? More productive and inspired? Have more time to spend on expressing your vision?

Sounds cool, right?

But what if by doing this you have increased demand placed on you to deliver more?

What if, by automating mundane tasks, entry level jobs dry up, so those avenues to learn the basics of the craft disappear?

What if the content that gets created ends up being more derivative, rather than innovative, as it is being generated not by people but by AI-enabled engines?

Doesn’t sound quite so appealing, does it?

The reality is that these types of issues – and many more – are what creative companies such as Avid are wrestling with. Avid AI developments incorporate AI advances to deliver practical benefits both to application users and to companies who want to ensure they are working and delivering in the most efficient way possible.

In a recent European Broadcasting Union (EBU) event in Geneva, Switzerland, Avid’s Régis André, Senior Director of Product Management, outlined how Avid is researching AI technologies and seeking to implement them within their products and solutions.

Régis André, Senior Director of Product Management at Avid

The annual EBU Production Technologies Seminar, Pioneering Tomorrow’s Tech Today, saw three days of in-depth presentation and discussion on an array of topics concerning the media production supply chain. From discussions on transformative technologies and the future of production to the use of large language models (LLMs) in news and utilizing the cloud, the event brought together leading technologists from EBU member organizations and top technology companies to look ahead to the rest of the decade and beyond.

The session André presented was entitled “Fostering Creativity Through the Power of AI.” In 2023, Avid launched Avid Ada, their artistic digital assistant technology, which is being implemented in a variety of the company’s products and solutions. Some of these solutions were demonstrated as tech previews at the IBC Trade Show, in Amsterdam, and you can see a webinar highlighting them here.

In his presentation, part of the sessions entitled, “The Crystal Ball Session – Tech for Creatives in 2030,” André highlighted both the benefits and potential pitfalls of implementing AI within the media technology sector. While potential benefits include the freeing up of staff to focus on the more creative aspects of their work, the downsides of AI need to be considered, too. This is why Avid, along with other technology companies, is following guidelines around the use of responsible AI.

André explained the methodology which Avid is applying when it comes to researching and prototyping technological advancements around AI. The aim is to develop quickly, gain customer feedback through the use of technology previews or proofs of concept, and apply those learnings to further refine workflows and deliver final products.

Avid has already introduced AI into some of their products. In fact, AI has existed for some time in MediaCentral, where the use of cloud-hosted AI services, has been implemented to offer automated metadata enrichment through the use of facial recognition, optical character recognition (OCR), speech to text transcription, emotion detection, and more.

AI Search In MediaCentral

Avid is working on using Avid Ada as a co-pilot for the application user. In Avid’s view, the user is very much in charge, taking advantage of AI-enabled technologies to help them work better, but remaining in control.

In Avid’s industry leading non-linear editor (NLE), Media Composer, two implementations of AI were added in 2023– ScriptSync AI and PhraseFind AI. With ScriptSync AI, editors can automatically generate a transcript from an interview, and with PhraseFind AI, editors can search for words or phrases in interviews and quickly edit into their timelines.

But André pointed out that these implementations are only the start, and Avid is working on extending these capabilities. He shared one tech preview, where the script generated from an interview could quickly be turned into a subtitle file and identify the speakers as well, potentially saving hours of work.

ScriptSync AI

André talked about the technologies behind new methods of searching and metadata enhancement, using what are known as semantic technologies and knowledge graphs.

He explained that semantic technology with knowledge graphs, build relationships between data in various formats and sources, to help create context. When they are linked together in this way, the various pieces create a giant web of data which can be interrogated.

IBM describes a knowledge graph “as a semantic network, (which) represents a network of real-world entities—i.e. objects, events, situations, or concepts—and illustrates the relationship between them. This information is usually stored in a graph database and visualized as a graph structure.

Applying these technologies to real world use cases from the media production industry, could be extremely beneficial to a variety of the creative industries, according to André. For example, in news, journalists could write their scripts for stories and at the same time be offered footage directly related to what they are typing (called A-roll), or background material (called B-roll) which would further enhance the way they tell the story.

Recommendation Engine

An example would be typing in a story about the United States president. The A-roll would provide shots of Joe Biden, the B-roll would offer shots of the White House or the Oval Office.

André described how this could transform the way people access and find archival material, again cutting the time needed to complete certain tasks. The industry is moving beyond asset management to “knowledge management.” This is a crucial shift. Knowledge management is described as the process of organizing, creating, using, and sharing collective knowledge within an organization. Successful knowledge management includes ensuring information is easy to access. By effectively implementing knowledge management – within products, solutions, and organization workflows – companies and processes can be transformed.

Avid is working across its product portfolio to take advantage of the advances seen in AI and machine learning, applying them to a full range of use cases to assess the potential value and practical application of the technologies. This assessment is not simply based on the technical details of the solution, but also around assessing workflow challenges the solution can solve.

To find out more about Avid AI developments, why not check out www.avid.com/ada