AI conversational agents like ChatGPT are evolving at breakneck speed. Shovel-ready speech and video integrations will bring AI agents to life as co-workers, advisors, and friends. The relationships with our newfound companions will contain three ingredients — personification, personalization, and perspective.
Feelings For Furby
In 2011, the famous Furby experiment demonstrated how easy it was to tap into human empathy. In the experiment, a group of children each received a Barbie doll, a live hamster, and an animated toy named Furby.
They asked the children to hold each of these three, upside down, one at a time. The children had no problem holding the inanimate Barbie inverted forever, whereas they chose either not to flip the hamster at all or only for a few seconds after seeing the little animal in distress.
Clearly, the children’s hearts went out to the living creature. But, when asked to do the same to Furby, it was different. The children knew the cute animatronic was a toy, but when it began to say, “I’m afraid.” Or “This is scary.” They quickly turned the toy right-side up.
When interviewed later, the children described a “feeling” they had for the cute, vulnerable Furby. As H. L. Mencken said, “Love is the triumph of imagination over intelligence.”
Our New Furby
Today, we welcome our new Furby, and it appears to be love at first sight. ChatGPT, the popular chatbot from OpenAI, reached 100 million monthly active users in January 2023, just two months after launch, making it the fastest-growing consumer app.
Despite reports of fabricated and outdated information, businesses and individuals are cautiously embracing ChatGPT to perform a myriad of tasks. ChatGPT has given the internet’s vast collection of human knowledge a voice, albeit text-to-text.
That voice has been heard around the world, with investors pouring billions into the creation of advanced chatbots. With many text-to-speech and AI video generators ready for integration, a session with your favorite chatbot will soon resemble a Zoom call with a new colleague — or Steve Jobs.
Personification: AI as Individuals
AI agents, like ChatGPT, are getting better at generating conversations that feel eerily human. When asked to demonstrate different personalities in response to the question, “What’s your favorite book?” ChatGPT replied:
Enthusiastic: “Oh my gosh, I absolutely LOVE ‘To Kill a Mockingbird’ by Harper Lee! It’s such a thought-provoking and beautifully written novel that always leaves me with so much to think about!”
Humorous: “My favorite book? Definitely ‘How to Be a Professional Couch Potato’ by L. A. Zee! Just kidding! But in all seriousness, I have a soft spot for ‘The Hitchhiker’s Guide to the Galaxy’ by Douglas Adams. It’s hilarious and always makes me laugh.”
Shy: “Um, well, I really like ‘Pride and Prejudice’ by Jane Austen. It’s a nice classic, and I enjoy the romance and social commentary. I hope that’s a good answer.”
In the future, AI agents will become more human-like and customizable. Imagine creating a new hire or companion from a diverse set of appearance, voice, personality, and cultural options.
Expect to engage with a menagerie of specialized AI agents trained on datasets specific to their role as personal assistants, educational tutors, historical characters, health advisors, or personal trainers. Each with its own persona befitting its role.
Despite Hollywood strikes, actors and celebrities will find it hard to resist licensing their personas for use in AI agents — with revenue streams flowing long after they have left the building. The next step will come as ‘regular’ people imprint themselves into AI digital clones to interact with generations yet born.
Personalization: AI Tailored for You
AI agents learn from every interaction, improving and adapting responses to better blend with a user’s persona. This personalization will make AI feel less like a tool and more like a personalized digital companion, increasing user engagement and trust.
Your personal AI assistant might learn you like waking up with a gentle alarm sound, and that you enjoy a summary of positive news with your morning coffee. Your AI assistant will adjust its interaction style, its recommendations, and its conversation topics to suit you individually.
An educational AI might adopt the persona of an encouraging tutor, making the learning experience more enjoyable for students. If an AI detects its user is feeling down from the tone of their messages, it might express empathy and provide uplifting responses, much like a friend.
In more intimate scenarios, AI avatars could act as life coaches, guiding users toward their personal goals. There’s even the potential for AI avatars to serve as romantic partners, though this raises significant ethical and privacy questions. Consider when your AI companion knows you better than anyone. Might it deliver your eulogy?
Perspective: AI With Worldviews
Although AI doesn’t have beliefs, feelings, or experiences, it doesn’t mean AI can’t simulate a perspective. AI models are trained on vast amounts of data, covering numerous beliefs and worldviews, allowing them to generate responses that mimic any perspective.
An AI interacting with a user who expresses environmental concerns might respond with information that aligns with an environmentally conscious perspective. Conversely, an AI trained on data representing a racist or bigoted ideology may reinforce those beliefs in the user.
The fear is we will repeat the recent past, where a fractured media landscape walled off viewers inside tribal echo chambers. The fracturing of AI has already begun, with Elon Musk announcing he will build an anti-woke AI agent to offset his perception of ChatGPT’s tendency towards political correctness.
Conversely, countries that wish to control their citizens will want to restrict usage to monolithic AI agents espousing only their leadership’s worldview. Will we be able to choose our AI agent’s perspective, or will it be chosen for us by a dictator or eccentric billionaire?
Beware The Gremlins
As AI becomes more human-like, there is a risk of people developing an emotional overdependence on AI, leading to erosion of human relationships, social skills and negative health outcomes, such as decreased physical activity, sleep deprivation, or mental health issues.
With increased personalization, there’s a significant concern about privacy. For AI to truly customize experiences, they must gather and analyze vast amounts of personal data. This could lead to an invasion of privacy if not handled ethically and transparently.
Agencies responding to calls for AI regulation need to consider the “law of the minimum and the maximum” to ensure AI agents don’t run amok, leading users down rabbit holes paved with misinformation while also preventing authoritarians from turning personal AI companions into the spy next door.
The Future is Personal
In conclusion, while personal AI agents can lead to improved user outcomes and economic benefits, it can also contribute to individual and societal issues. As these avatars become integral to our lives, we need to remain mindful of how they’re shaping our experiences, perspectives, and norms.
Thank you for taking the time to read this article. If you enjoyed it, give me a clap, a share, send a response or follow me on Medium. Let’s continue our journey of learning together!
Insightful concepts about the future facing us! It seems to be leading towards a blending of voice, image, movement, behavior, … among ourselves and our avatars. Multidimensional digitally enabled future relationships.