Computers' history of processing voice commands dates back decades. But recent advances in artificial intelligence and machine learning mean this could be a watershed year for voice recognition technology.
Gathering May 10 at TV Connect in London, the broadcast and over-the-top communities discussed the challenges and opportunities associated with bringing voice control to the table, including new ways for users to interact with search, discovery and recommendations to ensure more diversity in content consumption.
Until recently, discovering new content required users to pore over large catalogs of content. But voice recognition and AI are helping to simplify the vast world of content across a multiple devices and content platforms, according to Sylvain Thevenot, managing director of entertainment software firm Netgem SA.
"The double challenge we face is … the huge amount of content sources, [including] linear, on-demand, catch-up and OTT, as well as the different audiences which have different behaviors and different expectations about how discovery should work," Thevenot said during a panel discussion.
Voice control and the prospects for going beyond conversational interfaces go hand-in-hand with higher levels of smartphone penetration. In recent years, however, the tech industry invested heavily in AI and voice-driven assistants such as Apple Inc.'s Siri, Amazon.com Inc.'s Alexa, Microsoft Corp.'s Cortana and Alphabet Inc.'s Google Assistant, all fighting for dominance on users' smartphone screens, speakers, TVs, search bars and car dashboards.
Although the use of most voice technology remains limited to shortcuts and search commands for the now, Thevenot said that it had gone a long way toward streamlining the user experience.
But despite advances in voice control, panelists agreed that not even streaming juggernaut, Netflix Inc., has managed to crack the code when it comes to content recommendations.
"I am a big fan of Netflix but I see a problem with the algorithm [because] … an algorithm does not know which mood you're in," said Olivier Jollet, managing director of Pluto TV, an Internet-based TV platform.
As the industry shifts toward machine learning to provide better recommendations, providing a human touch would become equally as important.
"You need to use algorithms to improve the personal offering of each user but at the same time, it's important to have a strong editorial team and that's something that you can't get from a machine," Jollet said.
Adding to that, a number of localization challenges remain, particularly with processing different accents.
Noriko Matsuoka, lead strategist of emerging tech at the BBC, said that because most of the developments in voice have largely been trained on American software, there has been an issue with much of software not understanding U.K. or regional accents.
Voice control is "a lot more mature, more established" in the U.S. market, but tech companies are increasingly trying to localize and tailor their software to European and international users, Matsuoka added.
And given the staying-power of the remote control for linear TV audiences, the use of voice recognition in the TV industry has so far gained much more traction among online-only players.
Sion Wynn-Jones, product director at YouView, smart TV platform developed by U.K. broadcasters and telecoms groups BT Group, TalkTalk, Arqiva, the BBC, ITV PLC, Channel 4 (UK) and Channel 5 (UK), said traditional content providers have been somewhat cautious.
"Our shareholders are very scared about some capabilities … the public service broadcasters obviously enjoy prominence so how a disruptive technology like voice powered by AI, how is that going to impact them … does worry them," he said.