Is It OK to Use AI Avatars in Employee Communication?

Internal Communication in the Age of Artificial Intelligence by Monique Zytnik will be published by Business Expert Press on 30 May 2024.

In today’s rapidly advancing tech landscape, your CEO can send personalized messages to every employee, tailored and thoughtful. This may sound ideal, but there’s a twist: the CEO isn’t the one directly communicating. Instead, an AI avatar, a digital clone with eerily similar speech and appearance, delivers these messages. This scenario is not plucked from science fiction, an episode of Star Wars, but a reality made possible by advanced AI technologies specializing in synthetic media and video generation. Back in 2022, when I started writing my soon-to-be-released book, Internal Communication in the Age of Artificial Intelligence, I could not have imagined realistic, interactive AI avatars would already be a reality.

AI avatars, which can be trained to mimic human nuances in voice and appearance, are stepping into roles that were traditionally reserved for humans. As intriguing as it sounds, it prompts ethical and practical questions for employee communication: Is this technology ok to use in the workplace? What are the risks and benefits?

What are AI Avatars?

AI avatars are sophisticated programs designed using cutting-edge AI, particularly in the fields of natural language processing and machine learning, combined with synthetic media. These avatars are not just virtual representations but can interact in a human-like way, learning from interactions to improve over time. For example, tools developed by companies like Synthesia, DeepBrain AI, TrueBlueAI, and Nvidia allow organizations to create and deploy AI avatars that can give speeches or conduct training looking and sounding like real people, such as your CEO or even yourself. They are being used in external comms by companies such as Samsung Securities to create AI financial analysts and Roche to create AI doctors, both of which are trustworthy roles that interact with customers. They are more than our tailored CEO video, or favorite Keanu Reeves or Tom Cruise YouTube laugh because they go beyond just deepfake video content to be able to communicate in a dialogue with people like existing text-based chatbots.

📚Learn more: Best AI Prompts for Internal Communications [40+ Ideas]

• Trust concerns: Employees might feel a disconnect if they believe interactions are not genuine, potentially leading to diminished trust in company leadership. Once you’ve sent your personalized video and you know it is computer generated, is it so special after that? Senior analyst at Forrester, Rowan Curran, advocates rigorous testing to understand how employees respond to the AI avatars in action, in addition to security and governance testing.
• Privacy and security: The deployment of AI avatars involves handling vast amounts of personal data, making privacy and security paramount. The potential misuse of this technology to create deepfakes also raises concerns about the wrongful use of someone’s likeness and voice, further complicating the ethical landscape. You would need to be able to trust your employees generating content with access to the trained avatars. Deepfakes damaging brands should already be a part of your business continuity and crisis planning. Consider how you would play this if it came from an employee, intentionally or unintentionally.

Companies such as German supermarket chain REWE are already trialing ‘hologram’ versions of these AI avatars, allowing them to interact with employees and answer their questions. The REWE avatar is a digital clone of their Digital Manager and stands in a larger-than-life, lightbox-like structure. It is almost as if you are looking through a doorway and interacting with another employee. REWE is reporting that employees like to be able to ask some questions anonymously (such as the number of sick days remaining or childcare options). But my question is whether it is truly anonymous if equipped with cameras and audio recognition. As we delve deeper into the possibilities, the core question arises—while the technology is here, should we use it? What are the pros, cons, and considerations?

🧐 Check out Trust but Verify: Navigating the Age of AI

Let’s Start With the Positive

Although ChatGPT would like me to tell you that in various sectors, the adoption of AI avatars has been shown to boost efficiency and streamline communication, particularly in large or global companies, I can’t put my hand on my heart and say this is true. Having worked in a SaaS scale-up and with tech companies, there is a degree of “smoke and mirrors” that surround “product news.” Plenty of speculation, and lots of fabrication. We all know that proper piloting, trialing, and evaluation with real people, in real commercial environments takes months. Tried and tested is best. So here I will summarize the possible benefits:

• Provide 24/7 availability: Your CEO can always deliver no matter what time of the day. Unlike human counterparts, AI does not need rest, making it great for continuous employee support, especially in organizations spanning multiple time zones. According to a survey of 8,370 employees, managers, and HR leaders across 10 countries by Oracle, 64% of people would trust a robot more than their manager for advice, and the same study indicated that AI advisors are preferred for their constant availability. 
• Consistent messaging: AI avatars could help in delivering consistent messages across the entire organization. AI ensures that every employee, regardless of location or department, receives the same information in the right tone of voice (once you’ve trained your avatar of course).
• Scalability: As businesses expand, maintaining effective communication across growing teams and new geographies can be challenging. Going back to our customized CEO video messages, is it really any different from having your text-based newsletter messages from the CEO start with your name?
• Easily updatable: Type of voice, appearance, and content can easily be updated.

💡Want more insights? Check out Top 5 Ways AI Can Boost Internal Communications

Potential Challenges and Ethics

The first one that springs to mind is cost, however, there are others, equally, if not more important. Despite the apparent benefits, the use of AI avatars in employee communications is not without its challenges:

• You need the budget: Any tech implementation needs a proper project plan, budget and accountability. For example, are your IT systems robust enough to cope with the additional data storage needs of personalized videos to every employee?
• Guidelines: You need a clear understanding of who is allowed to be “avatared” (I just created a new verb) and for what purpose. The guidelines need to be robust and fit with your existing employee, AI use, and other relevant guidelines. For example, the fashion industry is already facing a battle with further unrealistic human body stereotyping through AI-generated models. How realistic (in a flawed human sense) should your avatar be? Is it inclusive? 
• Misinformation risk: AI systems, despite their sophistication, can still err, especially in complex, context-dependent situations. The dissemination of incorrect information, even unintentionally, can have severe repercussions for trust and compliance within an organization, particularly if it is delivered with the face and voice of someone with authority. A human must ultimately be responsible for the content and safety.
• Regulation compliance: European AI Act that was passed on 13 March 2024, emphasizes transparency and accountability in AI applications, mandating clear labelling of AI-generated communications. This regulation aims to ensure that employees are always aware when they are interacting with an AI, not a human and protect privacy.

While the above list is a bit of a riff on the blues of generative AI, there is still a lot of possibility for new gremlins and Clone Wars pirates to emerge and give us fresh considerations. I also caution against running towards the shiny new tech without working through the actual communication needs of your people and company. Have you got the basics right? Can you actually reach your people?

🚀 Haiilo’s Content Creation goes AI: Introducing Haiilo AVA

Where to From Here?

As synthetic media becomes more prevalent, my money is on us celebrating the imperfections that we find in humanity. These technologies allow us to become hyper-personalized and focus on the experience, but we must provide a great experience. Great technology will not solve a poorly thought-out communication campaign or fix a half-baked project. And, you cannot roll out a new technology project without an adoption plan. There needs to be a clear purpose and benefit.

To ignore the benefits that technology offers is also a folly. If you have the budget, any budget, it is time to play with what is possible. In a safe, learn-by-doing environment, see what works for your people and company.

What are your thoughts on integrating AI avatars into your communication strategies? Have you experienced AI in your workplace, and what were the outcomes? Join the conversation on LinkedIn to continue this discussion.

Internal Communication in the Age of Artificial Intelligence by Monique Zytnik will be published by Business Expert Press on 30 May 2024.

📖 Continue Reading: Generative AI in Internal Comms: Your New Intern for Better Communication

This blog article was co-authored by a human and ChatGPT.

About the Author

Monique Zytnik, IABC EMENA Region incoming Chair 2024/25

Monique Zytnik is an award-winning, global internal communication leader based in Berlin, Germany.

She’s worked internationally, presented on best-practice communication at world conferences, and guest lectured at universities. Her campaigns have been recognized by Gartner, Mumbrella CommsCon, and the Digital Communication Awards. She regularly shares her knowledge through communication industry publications and podcasts, drawing on her in-house and consultancy work with organizations including eBay, DHL Group, The Australian Government, SBS Radio Australia, and the LEGO Foundation.

References

Deepbrain AI (2024) Roche transforms healthcare with AI innovation, enhancing patient care and accessibility. https://www.deepbrain.io/customers/roche

Finnegan, M. (25 January 2024) The rise of synthetic media: Get ready for AI avatars at work. Computerworld. https://www.computerworld.com/article/1611603/the-rise-of-synthetic-media-get-ready-for-ai-avatars-at-work.html 

Oracle. (15 October 2019) 64% of People Trust a Robot More Than Their Manager https://go.oracle.com/LP=86149

Pinhammer, J. (18 April 2024) Rewe pilots use of avatars in employee communication. Retail Optimizer. https://retail-optimiser.de/en/rewe-pilots-use-of-avatars-in-employee-communication

Scott, L. and Armstrong, R. (5 May 2024) How modelling agent Chelsea Bonner is taking on artificial intelligence to avoid ‘cataclysmic step backwards’ for industry. ABC News.  https://www.abc.net.au/news/2024-05-06/chelsea-bonner-fight-against-ai-models/103769766

Synthesia (2024) https://www.synthesia.io

Tan, A. (22 April 2024) How businesses can embrace a phygital future with digital avatars. Ernst & Young Advisory. https://www.ey.com/en_ph/consulting/how-businesses-can-embrace-a-phygital-future-with-digital-avatars

Tsymbal. O. (16 March 2024) The Rise of AI Avatars in Business. European Business Review. https://www.europeanbusinessreview.com/the-rise-of-ai-avatars-in-business/

Unreal Keanu Reeves (2024) https://www.youtube.com/@unreal_keanu