From Hype to Reality: Avid Ada and the Future of News

One year ago, we were in the white-hot heat of an AI summer. Anything seemed possible. ChatGPT had opened the Pandora’s box of what AI and machine learning could do, and the world of journalism and media production was changing forever. In fact, never mind this industry, the world itself was going to be revolutionized, prompting fevered discussion on what this meant for society.

Spin forward twelve months, and while the hype of last year has receded somewhat, news organizations, broadcasters and the wider media industry continue to grapple with what this type of technology means today and in the near future.

At the recent IBC2024 trade show in Amsterdam, Avid delivered a session called, “From Hype to Reality: Avid Ada and the Future of News,” outlining practical ways in which we are rolling out AI across our product portfolio. But the presentation also focused on innovative ongoing research work which we are engaged in to deliver new workflows in media production by taking advantage of what is described as “knowledge management.”

The presentation was given by Craig Wilson, Avid Product Evangelist, and Régis André, Senior Director of Product Management, as part of the Skybox Session presentations at the show.

André and Wilson introduced Avid Ada, a framework of technologies upon which all Avid AI-related solutions will be built. This is the foundation for Avid’s AI-enabled future.

Avid Ada is already part of Avid’s industry-leading, non-linear editing application, Media Composer, with ScriptSync and PhraseFind, and a new transcript window. At the trade show, similar transcription capabilities were shown in MediaCentral | Cloud UX, a browser-based application which powers a range of workflows around news and general media production. This will be included in a forthcoming release of MediaCentral. And as a tech preview, transcription was also shown in Pro Tools, Avid’s award winning Digital Audio Workstation (DAW). 

“All of these AI implementations are running locally,” Wilson says. “There is no need for any connection to a cloud-hosted service, and this is an important aspect of AI, the requirement to ensure protection of content and any customer’s intellectual property (IP).”

Transcription services are only one element of AI, important and foundational, as other services will be built on top of them.

André outlined innovative research work which Avid is undertaking to bring new AI capabilities to the market, eliminating inefficiencies in workflows and duplication of effort, and enhancing the ability of any member of the editorial team to both understand what they have, and to find it faster.

He highlighted the DIKW Pyramid methodology to define knowledge. It starts with data, moves on to information, then knowledge and then wisdom and it is wisdom which people bring to the equation.

“This is what makes us different and what offers the opportunity to stand out from the realm of AI, particularly when it comes to original thought and content creation,” André explains.

Andre added that there are many challenges facing journalists when it comes to sources of information, and that today, some sources are disparate and scattered, even within one news organization. For example, there are text wires from news agencies, stories which are written for digital and broadcast, videos captured by camera crews and satellite sources, audio from interviews and so on, that may all be stored on separate asset management systems.

Avid’s Research and Advanced Development (RADLab) is looking into how to use Large Language Models (LLMs) with Retrieval Augmented Generation (RAG) to begin to build relationships between the information stored within these systems. Taking advantage of other technologies such as knowledge graphs and semantic embeddings, the RADLab has created a proof of concept where a knowledge management system can make sense of the various sources of data, linking them together and extracting meaning from them.

From an end user perspective, journalists, for example, could then use this system to perform tasks such as locating wire stories, understand the context of stories and how they relate to each other, and even find related video footage. All of this can happen in a matter of seconds, be done in their own language by interrogating a chatbot for instance, enabling the journalist to focus on their skillset to create.

“The move from the traditional asset management systems of today to the knowledge management systems of tomorrow stands to be one of the biggest shifts in how the media industry works.” André explains. “We believe this work has enormous potential to deliver true benefits, both in terms of creativity and efficiency.”

The challenge for journalists today is to prove that they are better than the AI. It is their challenge to come up with the original question an interviewee has never been asked before, or to consider a story from a different angle. AI by its very nature is derivative, based on existing information. Journalists are the ones who bring wisdom to the task.

“It was once said that the job of a journalist could be summed up like this,” Wilson says. “If one person tells you it is raining, and another person tells you it is not, it is not the job of the journalist to report both opinions. It is the job of the journalist to look outside and see what the weather is and report the facts.”

By delivering innovative solutions which give journalists more time to focus on key parts of the story, establish the facts and then share that information with as many people as possible in the shortest space of time, then that wisdom which we have mentioned can be appreciated fully.

To find out more about Avid AI and the future of news, check out www.avid.com/avid-ada