We Made An AI Video And Our Avatar Went Rogue – OpEd

Shouldn’t a magazine about the possibilities of artificial intelligence actually use AI to create some content? That’s the conversation we had at IESE Business School Insight magazine in January 2023 when we published AI: Using your power for good.

Working with our Art Directors — the Madrid-based design agency Prodigioso Volcán — we prepared a script based on the infographic they had created for us, Train your brain for AI (available for download at the end of this article).

Prodigioso Volcán used the platform Synthesia, which was founded by a team of AI researchers and entrepreneurs from University College London (UCL), Stanford, Technical University of Munich (TUM) and Cambridge “to radically change the process of content creation and unleash human creativity for good” and to make sure that such “synthetic media progresses in an impactful and ethical way,” according to the Synthesia website.

There, anyone can choose from among 85+ avatars and 120+ voice options with different looks, languages and accents. Prodigioso Volcán sent us an audition reel so we could select among various options, in English and Spanish. From that, we chose our two presenters, Anna and Jamie.

Our script was entered into the platform. With Prodigioso Volcán designers, we had to indicate where to fine-tune the mouth movements to go with the voices, adding the stresses, pauses and intonation in the right place, to make them sound as natural as possible. In doing so, the platform users are effectively training the AI to sound more human-like, with better- and better-quality delivery.

The result is still not 100% perfect (as you’ll see from our video). But how long before such videos are so lifelike that most viewers won’t even notice?

A double-edged sword

Apart from being a fun experiment, this video highlights the double-edged nature of AI — a point which was raised in our report. Using avatars for an educational clip, like ours, is indeed cost-and-time efficient: One can achieve satisfactory, professional results without always needing to hire expensive actors, voiceover artists, equipment and studios, investing in complex video editing tools, or needing specialist knowledge. However, those same “benefits” imply a hollowing out and eventual redundancy of all those roles, tools, knowledge and skills. This spells major disruption.

Additionally, not all users of this AI have the best of intentions. We were surprised one morning in February to see “our” Anna in the news. She was being used to spread misinformation via bot accounts on Twitter and Facebook. This is a breach of Synthesia’s terms of service. But there she was, being exposed on CNN as a “deepfake.”

All of which serves as a reminder that “where AI goes next is ultimately down to its human users,” in the words of our report. “There is power, and there is good, and they need to come together,” IESE Dean Franz Heukamp stated at IESE’s 2022 Global Alumni Reunion (GAR) in Munich. And as business leaders “explore the great potential AI has for products, processes and services to improve people’s lives,” they must also “advocate for the responsible use of AI, which preserves the dignity and freedom of all human beings.”

Or as IESE professor Javier Zamora put it: “If we want AI for good, then it depends on us to make it so.”

We at IESE Business School Insight will continue to report on these issues, bringing the latest thinking, knowledge, research and best practice to our readers, always in accordance with IESE’s mission to put humans and ethics at the center, and always encouraging business leaders to — like the GAR strapline says — use their power for good.

Any thoughts on this topic? Please let us know at [email protected]

We’d love to hear from you!