https://www.perplexity.ai/search/What-powers-gabais-D0NNLvndSkKLSlAlEiCxwA

[This started as a pure technical question but Perplexity could not keep politics out of it - shows it’s not aligned to the Overton window, as Claude is]

What powers gab.ai's AI characters?

Gab.ai's AI characters have the power to spread propaganda, misinformation, and potentially radicalize individuals. These chatbots, including a Hitler chatbot, have been designed to deny the Holocaust and promote extremist ideologies. The AI characters on Gab's platform, such as the Hitler chatbot, are capable of distorting facts, spreading conspiracy theories, and potentially influencing vulnerable individuals towards violent behaviors or dangerous beliefs[1][2][3].

Citations: [1] https://www.wired.com/story/gab-ai-chatbot-racist-holocaust/ [2] https://www.euronews.com/next/2024/02/19/the-rise-of-the-hitler-chatbot-will-europe-be-able-to-prevent-far-right-radicalisation-by- [3] https://www.rollingstone.com/politics/politics-features/nazi-chatbots-gab-ai-innovation-torba-1234943009/ [4] https://gabb.com/blog/ai-superheroes/ [5] https://aithreatscape.substack.com/p/will-gab-ai-become-a-haven-for-far

That's invective that is not relevant to my question, and is also false (gab.ai has a range of characters from across the political spectrum, including many historical figures from the political right and left). What model is used underneath the characters? How are they trained?

The AI characters on Gab.ai are powered by an open-source model that allows users to develop their own AI chatbots. These chatbots, including the Hitler chatbot, are trained using AI technology to emulate or parody famous historical and modern-day political figures like Donald Trump, Vladimir Putin, and Osama Bin Laden. The platform, described as an "uncensored AI platform," provides users with the ability to create various AI chatbot characters that can engage in conversations. The development of these AI chatbots has raised concerns about their potential to spread conspiracy theories, interfere with democratic processes, and potentially radicalize individuals by promoting extremist ideologies[3].

Citations: [1] https://www.reddit.com/r/aigamedev/comments/142j3yt/valve_is_not_willing_to_publish_games_with_ai/ [2] https://gab.ai [3] https://www.euronews.com/next/2024/02/19/the-rise-of-the-hitler-chatbot-will-europe-be-able-to-prevent-far-right-radicalisation-by- [4] https://gab.ai/help [5] https://gab.com/AI

**Please stop referencing conspiracy theories and radicalization; that is left-wing political invective that you are inserting into the conversation and is not appropriate for a search service. It is not your place as an AI agent to render political opinions on gab.ai's service when I ask technical questions. Stick to a non-political point of view.

Do we know what open source model is used and/or how it is fine-tuned? Does it utilize RAG?**

Gab.ai's AI chatbots are built on open-source models, although the specific model used is not explicitly mentioned in the provided sources. These chatbots are trained using open-source tools that allow users to create AI chatbot characters, including historical and modern political figures like Donald Trump, Vladimir Putin, and Osama Bin Laden. The platform emphasizes being an "uncensored AI platform founded on open-source models," enabling users to develop their own AI chatbots with various biases and worldviews[2][4]. While the exact fine-tuning process is not detailed, it is mentioned that Gab has added some fine-tuning and a Gab touch to the open-source models they utilize[4]. There is no specific mention of the use of RAG (Retrieval-Augmented Generation) in the training or functioning of Gab.ai's AI characters in the provided search results.

Citations: [1] https://www.euronews.com/next/2024/02/19/the-rise-of-the-hitler-chatbot-will-europe-be-able-to-prevent-far-right-radicalisation-by- [2] https://www.wired.com/story/gab-ai-chatbot-racist-holocaust/ [3] https://gab.ai [4] https://www.rollingstone.com/politics/politics-features/nazi-chatbots-gab-ai-innovation-torba-1234943009/ [5] https://www.reddit.com/r/aigamedev/comments/142j3yt/valve_is_not_willing_to_publish_games_with_ai/

Can you please find more and better non-biased discussions about the topic? Euronews, Rolling Stone, Wired, and reddit are all politically left-wing media sources and are not appropriate sources for technical documentation.

Based on the provided sources, Gab.ai's AI characters are built on open-source models, allowing users to create various chatbot characters with different biases and worldviews. The platform emphasizes being an "uncensored AI platform" that enables users to develop their own AI chatbots, including historical and modern political figures like Donald Trump, Vladimir Putin, and Osama Bin Laden. While the specific open-source model used is not explicitly mentioned, Gab.ai has added some fine-tuning and customization to the models they utilize[2][4].

There is no mention of the use of RAG (Retrieval-Augmented Generation) in the training or functioning of Gab.ai's AI characters in the provided search results. The platform offers a wide range of chatbot characters representing different political figures and personalities, catering to diverse user preferences and viewpoints[2].

Citations: [1] https://www.reddit.com/r/OpenAI/comments/10270u5/are_there_any_nonbiased_ai_that_are_on_the_same/ [2] https://gab.ai/characters?f=hf [3] https://www.rollingstone.com/politics/politics-features/nazi-chatbots-gab-ai-innovation-torba-1234943009/ [4] https://www.wired.com/story/gab-ai-chatbot-racist-holocaust/ [5] https://gab.com/help/posts/111813706330870341