Bridging Human Authenticity and AI

Posted on by

In a world where innovations shape our daily interactions and experiences, one aspect remains timeless: the way we consume news. As information becomes increasingly abundant and accessible, news organizations face the challenge of capturing and retaining audience attention in a landscape saturated with content. However, amidst this digital revolution, a new frontier emerges: personalized news delivery driven by artificial intelligence (AI) and virtual human presenters.

Enter Dr. Leon Hawthorne, a visionary researcher with a background in news anchoring for prominent media outlets, including CNN and CNBC. Dr. Hawthorne’s journey into the realm of AI-driven news delivery began with a simple question, How can we adapt traditional news formats to resonate with modern audiences?

With this question in mind, Dr. Hawthorne embarked on a groundbreaking research project titled “Talking Heads: The Use of Virtual Human Presenters for the Delivery of News Content.” This initiative aimed to evaluate the integration of AI-generated virtual humans as on-screen news presenters, exploring their impact on audience engagement, perceptions, and preferences.

I joined him to delve into the nuances of human behavior and the intricacies of AI technology. Together, we conducted extensive studies, analyzing audience reactions to various presenter models and news stories.

Do Audiences Prefer Real People or Virtual Humans?

Our findings unveiled a fascinating dichotomy: while audiences overwhelmingly preferred real human presenters, we also saw a nuanced acceptance of AI-generated virtual humans in specific contexts. This research highlights the delicate interplay between technological innovation, audience perceptions, and the future of digital communication.

But our journey didn’t end there. Drawing from Dr. Hawthorne’s insights into personalized news delivery, he envisioned a future where AI-driven virtual presenters could revolutionize the way we consume information. By harnessing the power of AI and machine learning algorithms, news organizations could tailor content to individual preferences, delivering personalized news experiences on a massive scale.

Technical Insights: Key Research Takeaways

Our research explored AI-generated virtual presenters, examining factors such as lip synchronization, facial expressions, and overall realism. Through a series of experiments involving different presenter models and news stories, the research revealed key insights into audience preferences and emotional responses.

Overwhelmingly, participants preferred the real human presenter, finding him most engaging, knowledgeable, and trustworthy based on the survey answers. The three non-human models were judged below ‘Average’. Yet virtual avatars generated stronger emotional engagement at an intuitive level based on the facial expressions from the audience.

Cartoon-like virtual models elicited, by far, the highest emotional positivity and generated the highest emotional engagement. The use of oversized babyish features in this model was calculated to create “supernormal stimuli” (Costa & Corazza, 2006) in order to elicit strong positive emotional responses. It succeeded in achieving this at the intuitive level, but the deliberative responses showed participants judged these features totally inappropriate for news, especially hard news stories about war and death. Hence the facial expression of a smile reflects how laughable and inappropriate this avatar is for the type of content delivered.

Moreover, the research explored the concept of “uncanniness” in virtual human design—a fine line between realism and artificiality. The virtual human avatar scored high on emotional engagement, but not in a positive way. This demonstrated its aesthetic engendered strong emotions, but these were predominantly negative emotions. Participants also showed high levels of ‘fear’ at the sight of this avatar, suggesting they found the design uncanny. Another notable finding was the importance of lip synchronization in creating a convincing virtual presenter. A discovery of this insight highlights the critical role of realistic animation in fostering viewer engagement.

Through rigorous facial coding data analysis via Affectiva’s Emotion AI technology, we uncovered intriguing data findings that shed light on audience perceptions. While real human anchors remained preferred for their authenticity, AI-generated virtual anchors showed promise in delivering non-serious news content with engaging narratives.

This phenomenon underscored the need for careful design considerations to mitigate uncanny feeling responses and enhance audience acceptance.

Where do we go? 

For AI and data specialists, this research offers valuable insights into the transformative potential of AI-driven content delivery, such as our daily news. By understanding the technical nuances of virtual presenter design and audience preferences, IT pros play a pivotal role in implementing AI-driven solutions within their organizations.

From developing advanced animation algorithms to optimizing user interfaces for personalized news delivery, tech professionals can leverage our findings to drive innovation and enhance audience engagement. By embracing AI-driven solutions, news organizations will stay ahead of the curve and deliver unparalleled news experiences to their audiences.

In the digital age, where information is abundant and attention is scarce, Dr. Hawthorne’s vision of personalized news delivery represents a beacon of hope for news organizations and audiences alike. By embracing AI-driven solutions and incorporating technical insights into their strategies, IT teams can help usher in a new era of media innovation, where every individual can access the news they need, when they need it, in a format that resonates with them.

Sasha Mukhanova is a seasoned researcher with 15+ years experience in consumer insights. With a robust background in servicing international clients across diverse industries like FMCG, retail, insurance, and automotive, she has honed her expertise in regional research programs spanning Russia, Europe, and MENA. Sasha’s passion for advertising and consumer behavior led her to serve as the former creative domain lead for the Middle East at Kantar. Renowned as the ultimate advertising enthusiast, Sasha’s keen interest in neuroscience and neuroeconomics drives her quest to unravel the ‘brain code’ and humanize technology. Currently, she contributes her expertise to Affectiva’s global mission with fervor and dedication.