The implementation of gen AI seems to get weirder and weirder, while we’re continually told about how valuable and useful such experiences will become.
Which could be true, I guess, but it is a little strange to see the pathways we’re taking to increased value and relevance.
Case in point, today, Google has launched a new AI experiment called “Portraits,” which is essentially AI-generated depictions of famous people, that can offer conversational advice and notes in the voice of the person they’re based on.
As explained by Google:
“Portraits, a new experiment from Google Labs, lets you interact conversationally with AI representations of trusted experts, built in partnership with the experts themselves. Our first Portrait features Kim Scott, bestselling author of “Radical Candor,” and offers AI-powered coaching inspired by her renowned communication and leadership principles.”
Coaching, from an automated system based on machine learning prompts. Sure, what could possibly go wrong?
“Imagine having direct access to Kim’s guidance to navigate tough workplace situations or practice difficult conversations. Simply ask your questions, and the Portrait will use Gemini’s understanding and reasoning capabilities to generate a relevant and insightful response, drawing directly from Kim’s content and speaking in her voice via an illustrated avatar.”
So it’s really just another way to find relevant segments of advice from this person based on conversational prompts. So rather than having to read all of Scott’s books, and/or memorize key approaches, the system will give you easy access to relevant sections, based on Gemini’s understanding of your query.
Is that beneficial?
I mean, maybe speaking with the actual person would offer some assurance and guidance, but relying on an AI system for the same seems like a far less valuable experience.
But then again, if this process can offer similar advice, based on the available corpus of knowledge created by this person, that’s maybe, kind of the same thing?
Maybe?
Google says that the unique value of Portraits is that they are directly linked to real-world trusted experts, like Scott, who have provided their own knowledge and voice.
“The foundation is the creator’s authentic content, which ensures conversations stay focused on their specific areas of expertise. We’ve conducted extensive testing and implemented user feedback mechanisms to proactively identify and address potential problematic scenarios.”
I would still presume that many of these experts would prefer to have real-world context for any advice they give, and I doubt the advice provided will be wholly relevant.
But maybe, as a self-help tool, with expert guidance, it could be another way to utilize AI systems for good.
Meta’s working on similar, with AI versions of creators that can respond to fans.
Meta’s hoping that this will provide additional engagement opportunities, without imposing additional workload on creators, but I don’t see why this would be valuable, as you’re not actually talking to the real person.
So why the facade? Why bother putting a face to an AI system when it is, in fact, just the same as talking to a bot?
Portraits is a little different, in that it is trained on each subject’s responses and output. But it still feels like it’s missing the key element of celebrity interaction, being the actual interaction and connection itself.
Either way, it’s another experiment, and it’ll be interesting to see how users respond to AI versions of high-profile folk.
Google says that Portraits is now available in the U.S., while people can also register their interest in their own Portrait depiction.
