How to Make Virtual, Humanoid Newsreaders a Good Thing
Virtual humanoids and artificial intelligence (AI) are increasingly popular technologies in the news industry. They open up for new ways of curating and selecting information and how it is presented. If done right, it gives individual users the power to decide what they want to be informed about and how and when. If done wrong…well, things can get very problematic.
The first robot news presenter on a national channel got its trial run in Japan back in 2014. However, it was in neighboring China that we saw the most recent developments. In 2018, the media Xinhua introduced two virtual human newscasters. One for their English and one for their Chinese broadcasts.
The upsides are self-evident. For one thing, virtual newscasters can work 24 hours a day. They don’t ever misread, and they stay objectively to what’s in front of them. Some of those strengths can also be flaws, but that’s something we will return to later.
Light in Rough Reception
On Chinese social media platforms, the new virtual human anchors got a rough reception. Some pointed out that their un-human behavior was off-putting, while others questioned what the move meant for journalistic values.
Without wanting to be too rough, you could say that the main issue was that the newscasters showed about the same level of warmth and personality as a standard fridge.
Part of the reason was that the newscasters were not, although advertised as such, genuinely AI-powered. Yes, they were using machine learning techniques to incorporate a real anchor’s likeness and voice patterns. However, in this case, that is a bit like saying that being able to mix colors correctly is the same as being able to paint beautiful artwork.
If the virtual news anchor wrote its script or wrote its own news stories, or perhaps reacted in a human-like way to features and segments, that would be a different story (pun intended).
AI Working as a Journalist
If we look at the world around us, AI systems are already producing news stories. In some instances, the AIs will write the initial report, and then a journalist will refine it. In others, an AI will write a story, an editor gives it a quick assessment, and then publishes it. In both cases, the AI systems enable media employees to focus on other articles, as well as more research and contact to sources, while simultaneously creating content and news that might be suited to more readers.
Similarly, machine learning is being deployed throughout many news organizations and through specialized apps, such as SmartNews, to personalize news content for individuals. These systems use algorithms to improve their accuracy gradually. One example of such a design comes from Neue Zürcher Zeitung. The media organization used AI to construct a flexible paywall personalized to individual readers based on hundreds of criteria. The move increased conversion rates by fivefold, thereby boosting direct media revenues. Furthermore, increased readership likely led to better opportunities for generating further revenue through advertising.
Getting to the Next Step
The abilities of these systems are impressive. Combining the use of AIs able to create, correlate, personalize and structure news and virtual humanoids capable of presenting the story would, in many ways, be a natural next step.
To quickly explain: Consumers of news (most of us) are already used to news feeds structured by AIs. We don’t always have flexible control over how, when — and by whom — the information is delivered to us. For example, we spend a lot of time in situations where existing systems don’t know if they should push updates or alerts.
If we integrated Virtual Human Agent (VHA) technology, users would have the ability to control and optimize their news feeds and have it delivered through human-like, two-way communication. User satisfaction and engagement would be increased, ultimately accelerating readers’ satisfaction and information levels and growing subscriptions for media companies.
VHAs as Your News Nexus
VHAs are flexible AI-powered systems capable of undertaking a wide variety of different job tasks and functions. Their humanoid shape and ability to understand and express various communication types, including natural language and non-verbal communication, make them excellently suited to work as the nexus for modern news delivery.
VHAs bring personality and connection to news delivery. Thanks to an in-depth understanding of their users, they can help predict what news a person would like to hear/see, and when the best time is to present it to them.
Furthermore, if companies and users had a place to build upon existing solutions in an open environment, we would create new solutions. One example of this would be AI that could correlate and make abstracts of ongoing issues by analyzing news from many sources. The Ludens Marketplace provides such an open platform for sharing ideas and developing new solutions.
How to Avoid the Pitfalls
One of the potential dangers of increasing the use of AI in news selection and presentation is that it could lead to increasing the ‘information bubbles’ that surround a lot of people today. We see things in the media that tend to respond to what we already believe. The same applies to social media platforms.
In both cases, part of the reason is that algorithms already influence what news we are presented with. If it thinks we like a specific type of information or news from a particular source, it will prioritize it. This can lead to a self-building cycle that leaves a user stuck in a bubble and not aware of — nor challenged by — other opinions than the ones they already have. Many see this as a significant reason for increased disharmony and division of societies around the world.
To avoid such issues being increased by using AI and VHA technology, systems may be equipped with a ‘challenge’ mode, where they, for example, look for sources of information outside your usual choices. This is just one suggestion. Thanks to the Ludens Marketplace, such a solution — and many others — can be created quite easily.
