AI Ate My Assets -- Video at 11

Robot newscaster reporting on Elon's SpaceX Falcon Heavy’s boosters

Right now, AI is being applied to your videos! While this may conjure images of Cameronian Arnolds laying waste to a post apocalyptic world filled with John Connor, his minions, and your videos -- it isn’t nearly that exciting. It isn’t even as exciting as a hyperventilated Muskivellian AI overload (sorry Elon), which is basically a description what happens when you turn democracy, and your videos, over to Facebook. Nope, this is applied AI and it is doing its job for you as I type.

In today’s big data environment companies often think that their algorithms are their greatest asset. By adopting this mentality, they lose focus on what they have actually been doing, what they have been creating. Algorithms are worthless without data to apply them to and companies have been throwing away 80% of their video data, according to Gartner, simply to save on storage costs. This is understandable if you think of these assets as having no lasting value. But they do have value – value that is unlocked by applying AI.

Much time has been spent in books exploring the ultimate AI, the singularity, but that piece of fiction, while grand and potentially scary, diminishes the actual results that applied AI can, today, bring to many different types of problems. Applied AI, even using our currently infantile machine learning algorithms, can add value to our existing assets in many different and important ways. It can transform that 80%, and all the energy and thought behind, into assets that have lasting value.

AI’s advances in Speech to Text made accessibility and searchability some of the first major enhancements to become production ready. Speech to Text with timecode metadata provides both accessible captions that are useful for the hearing impaired and in situations where audio levels prevent acceptable listening. As an added side effect of the captioning, videos become deeply searchable to the level of jumping to points in videos where specific words or concepts are being presented.

Face and object recognition is another area where the applied AI is ready for production. Much like the technology you are used to on your phone or computer for authentication, the recognition is not perfect. With training, however, it can easily recognize people who are participating in your videos and properly index their presence.

Further away are enhancements to the natural language processing of the AI. Intent, speaker mood and semantic chaptering are all features that are on the near horizon. Intent allows the AI to understand that, not only am I talking about cognitive video, I’m also writing a blog with the intent of informing about applied AI and its use in video. My mood was jovial, albeit a bit snarky, and this paragraph falls into the “Future Enhancements” chapter.

Inferno, Jolokia’s cognitive video platform, is riding the edge of AI and incorporating applied features as they become production ready. Because of Inferno’s strong live capabilities, the AI journey starts with a live broadcast which uses real time Speech to Text to create captions. Captions are then used to enhance accessibility and searchability of the videos in their on-demand format. Because the captions are available, Inferno uses AIs natural language processing and translation engines to create translated caption files for the on-demand videos allowing for quick distribution across many geographical and language zones.

As the technology progresses Inferno features will be added or enhanced. Indexing/Chaptering will gain an auto detect feature. Videos will automatically enhance their tag cloud/meta tags with automatically detected subjects, intents and moods. Speakers will be identified as part of the regular AI processing on the on-demand videos.

While an AI infested dystopia is a favorite theme, a more promising and nearer future of applied AI is unfolding and its benefits are becoming clear. Cognitive infused video platforms are here. Just be sure that you don’t leave them on after midnight and, whatever you do, don’t get them wet…