If anyone doubted the normalization of AI in the M&E space is a fait accompli, the IBC Show in Amsterdam put that notion to rest last month with the help of some new developments that promise to lower the threat of unwanted consequences.
One of the more significant cases in point involves the emergence of neuro-processing units (NPUs) as AI-optimized chipsets, which by supplementing CPUs and GPUs in network appliances and CPE could alleviate core AI processing workloads while addressing privacy rules that can restrict use case development. Notably, NPUs now implanted in some high-end set-top boxes offer a new perspective on how cable operators, telcos, DBS providers and even NextGen TV broadcasters with their ATSC 3.0 signal gateways might be battling with smart-TV OEMs and cloud-based super-scalers for whole-home dominance in the future.
An Evolution
For example, Vantiva, a company quietly operating under the radar since combining the home-networking units of Technicolor and CommScope, introduced an NPU-driven STB labeled ONYX, a far-field voice-controlled (FFV) device that officials said opens the door to next-gen AI-supported use cases. These start with things like identifying and locating specific events in video, sharpening resolution, decreasing film grain, and greater content personalization but will evolve to more functionalities in tandem with the development of home-oriented large-language models.
With immense AI data processing power, such devices have the potential to interact verbally at a personal level with everyone in the household by utilizing facial, voice and device recognition and compilations of past user experience while avoiding the privacy violations that come with shipping information to the cloud. “We’re working with our [MVPD] customers to explore applications that can be developed to help them drive user engagement and monetization,” Vantiva Chef Technology Officer Charles Cheevers said.
At the production end of the M&E service pipeline, the industry’s massive shift to cloud-orchestrated workflows across cloud-centric and hybrid implementations involving both premises- and cloud-based appliances has unleashed an explosion of AI-assisted solutions that are designed to automate a big share of the workload. But reliability is a big issue.
Indeed, the saturation AI branding across 14 IBC exhibit halls at a moment when many aspects of the technology have yet to meet industry reliability standards prompted one vendor CEO to joke it might be smart to convert his own company’s high-profile AI messaging to mean “Actual Intelligence.” To get beyond those dependability risks some vendors are going out of their way to provide verifiable assurance that their AI implementations pass muster.
For example, Telestream—which showcased a range of AI-powered solutions aimed at streamlining ingestion, content enhancement and delivery in customers’ media processing, production and post-production workflows—shares the level-of-confidence percentages scored by all company solutions using AI, according to Colleen Smith, Telestream’s senior vice president of product marketing and channel enablement. While reaping the benefits to the fullest extent possible, “we’re taking a cautious approach to AI,” she said.
The Great Equalizer?
There are many other issues beyond reliability, privacy and processing power consumption, not to mention the widely debated potential for job losses, that the industry has to deal with as AI saturates the ecosystem. For example, there’s a “great-equalizer” aspect to AI that threatens to undermine the competitive advantages various vendors have enjoyed, raising the question of “how all these organizations will survive,” noted Juan Martin, CTO and co-founder of streaming platform provider Quickplay.
“We’re a digital transformation company helping customers to build platforms and orchestrate solutions across a multiple partner ecosystem,” Martin said. As a gateway to the streaming marketplace, the company’s success depends on its ability to “sift through the options for our customers.”
At the show, Quickplay announced it has positioned its cloud-native content management, processing orchestration, generative-AI tools, dynamic advertising and other capabilities for access in the Amazon Web Services (AWS) Marketplace to address the industry’s need for “the smartest, fastest, most effective ways to engage and monetize viewers.” By offering its open-architecture approach to orchestrating what it deems best-of-breed solutions from a bevy of partners, the company is helping customers to operate in the OTT market with suppliers that go beyond reliance on AI algorithms in building well-conceived solutions, Martin said.
The Dark Side
On a much broader collaborative scale, another initiative aimed at sifting through solutions has launched as an IBC “accelerator project” targeting the dark side of AI, where deepfakes have become a threat to the broadcast news business.
“I don’t think there could be any more existential threat to media than we have in the form of misinformation, disinformation, manipulated images, fake images,” said Mark Smith, IBC Council chairman and head of the IBC Accelerator Program. “There’s a whole tsunami of content that’s coming at these trusted brands in our world of news broadcasters and news agencies.”
During a session moderated by Smith entitled “Design Your Weapons in the Fight Against Disinformation,” Tim Forrest, content editor at U.K.-based Independent Television News, noted, “The fakes are getting better. The technology to make them is improving, too, and it’s getting easier and easier to use.” He said “a qualified guess” is that there are about 34 million deepfake images, video and text messages spreading across the globe daily.
The new effort to shore up the ability of legitimate newscasters to counter the scourge was the brainchild of consultant Allan McLennan, president and CEO of the PADEM Group, and Anthony Guarino, until recently executive vice president of global production and studio technology at Paramount. As described by McLennan, the goal is to foster awareness of the deepfake threat and cooperation in doing something about it across the intensely competitive global news industry.
Six months since getting underway, some of the world’s biggest newscasters have joined the fight, including the Associated Press, CBS/Paramount, BBC, ITN, Globo and many more, McLennan said. “We’re creating a path to sharing information and addressing disinformation together,” he noted. Along with promoting greater awareness of the threat, the participants are sharing results from their experiments with technologies that can be used to flag disinformation so that it doesn’t get into the news pipelines, which is hard to do when everybody is competing to be first in breaking news.”
