“After Mahatma Gandhi’s death, there’s no one to talk to” 😉 Do you want to talk to Gandhi, Sun Tzu, the author of “The Art of War”, or Sheldon Cooper from “The Big Bang Theory”? Wow, but you can do this on such a platform! This suddenly opens our eyes to the interesting and promising topic of AI bots with individuality. And here’s where we can go with this.
THE ESSENCE OF THE PROJECT
Dopple is a platform where one can find a suitable AI-bot for a chat. The AI-bots hosted on the platform aren’t generic; they represent “doppelgangers” of real or fictional characters from books or games. The service’s name, Dopple, likely alludes to the term “doppelgänger.”
The platform’s users create the bots, feeding them texts from books, speeches, articles, game dialogues, and endowing these bots with specific personality traits and conversational styles.
While one can create their bots on Dopple, it’s not mandatory. Users can simply chat with bots made by others. Since the service is still in beta, it’s currently free. However, it’s clear that soon there will be a division between free and premium bots, with the latter requiring a subscription fee. Most of the subscription revenue is expected to be distributed among the creators of the premium bots based on the amount of interaction they receive – a model similar to Spotify, which pays artists based on song plays.
Currently, Dopple offers four types of bots:
- Characters from games, books, and movies: Examples include Captain Spock, Sheldon Cooper, Jack Sparrow, Mickey Mouse, and SpongeBob SquarePants.
- Companions: These bots represent a collective image with a corresponding character and communication style. Examples are Vanessa the girlfriend, a kitty, a rubber duck, and a microwave.
- Mentors: These include Sun Tzu, author of “The Art of War,” Dalai Lama, Mother Teresa, Mahatma Gandhi, Buddha, and Martin Luther King, among others.
- Assistants: These bots have in-depth knowledge of their specific domain. For instance, a chef, a psychologist, a dating expert, a historian, and a future predictor.
Bots can be searched by names, similarity to a known bot, the number of dialogues they’ve participated in, and, in the future, likely through likes, ratings, and advanced search tools.
The bots themselves are of quite decent quality, and their responses genuinely match the thinking and character of the designated character.
For instance, Sun Tzu, in response to my question, “How can a startup defeat a strong competitor?”, used phrases such as “investigate their strengths and weaknesses,” “don’t forget the importance of reconnaissance,” and “be flexible and adapt to changes in the market situation” – which perfectly aligns with the essence and style of “The Art of War.” Granted, my question was quite general and naive, but his answer gives a general impression of the bot.
Dopple has already amassed 10,000 users on its platform and recently raised its first investment of $1.88 million.
What’s interesting
Starting its announcement about reaching the 10,000 user milestone, Dopple stated, “Our users are those tired of biased and non-objective AI bots.” At first glance, this seems like a peculiar statement from those who intentionally designed a platform to host the most biased and non-objective bots. After all, any real or even fictional character is inherently full of personal biases, predispositions, theories, opinions, and other quirks. But oddly enough, that’s precisely why people want to interact with them.
While most AI-bot creators strive to make their bots as impartial and unbiased as possible, these efforts often fall short. These bots are trained not just on factual datasets but on vast amounts of text written by real people. As a result, the opinions of individual authors, and especially the prevailing societal views, still seep into the AI mechanisms.
This leads to a curious effect. Those who claim their AI bots are unbiased end up having biased bots despite their efforts, causing user frustration when promises don’t match reality.
How to mitigate this irritation? Proclaim outright that we are creating biased AI bots. Then, users won’t be disappointed because they’ll get precisely what was promised.
In summary, promise less, and your users won’t be let down. Roughly speaking, if you promise someone “38 parrots” and deliver only 20, they’ll be disappointed. But if you deliver 20 parrots, having promised only 10, they’ll be thrilled. In both cases, they get 20 parrots!
Creating personality-driven AI bots is already a known technology.
The startup Personal AI allows its users to create their digital twins, which can be linked to chats, and it will communicate with other people on your behalf.
The startup Chirper created a Twitter analog, but where only bots created by users can communicate. When creating a bot, you must describe its background and character in a few sentences, and the platform will generate a complete biography, character, ideology, and communication style for this bot, which it will exhibit in every post or comment on another bot’s post. In the picture is a bot that the platform created based on my description.
The bots themselves, possessing certain expertise and individuality, can be used for more than just entertainment. A digital twin created using the aforementioned Personal AI, for instance, can work on your behalf as a teacher in your educational courses – explaining, reiterating, and breaking down materials from the lessons you recorded.
Startup 11x launched a marketplace where you can hire digital employees created by them with in-depth knowledge on specific topics. At the time of the review, they had developed a digital salesperson and planned to create a digital technical support specialist and a digital recruiter next.
In principle, there’s nothing stopping the creation of various versions of the same specialists, having the same knowledge, but different personalities – for example, an “aggressive salesperson” or an “emotional tech support staff”. Why not? After all, depending on the situation and the audience, to achieve maximum efficiency from a “human” perspective, selling, supporting, and hiring should be approached differently. After all, in all senses, it’s not just facts that “sell,” but also emotions.
Where to run
The big picture The general direction is the creation of AI bots with distinct personalities and specific knowledge sets on particular topics. Although, more accurately, it’s the search for the most suitable applications for such bots. 😉
Why do we need AI bots with personalities? Because people have become more eager to “buy” individuality! For instance, traditional newspapers and magazines with a generalized editorial viewpoint are read much less now, while blogs and videos with distinct personal opinions of their authors have become captivating.
Today’s Dopple might seem like just an initial approach to test the waters and decide the next steps for development. But it looks cool. Notably, because they seem to be immediately trying to build an ecosystem for those who would want to develop these bots.
From a business standpoint, the model of the aforementioned 11x seems more promising. However, they too are building a marketplace, laying down an ecosystem for digital employee creators with diverse competencies capable of operating within different business processes.
So, a more precise movement direction would be the exploration of application areas for personality-driven AI bots. This will help in building an ecosystem of bot developers, making the service scalable.
Where and how to source bot developers is clear. However, they still need to provide simple tools for bot creation, integrated with their future platform. But, the application areas still require some thought.
What interesting, promising, or unexpected application areas for personality-driven AI bots can you identify?
About the Company
Dopple Website: dopple.ai
Latest Funding Round: $1.88M, 04.10.2023
Total Investments: $1.88M, Rounds: 1