# Five interesting categories of AI proto-agents already in OpenAI’s new GPT Store

Author: Fernando Nieto Lobato
Original publication: 2024-01-17
Spanish original: https://estrategiabyaleph.substack.com/p/estrategia-16-la-llegada-de-los-primeros
English URL: https://elcontemplador.github.io/estrategia-english/essays/016/
Status: Published translation

This is a translation of the original Spanish essay published on 17 January 2024. Its claims, examples and forecasts retain that historical context.

English publication: 2026-09-29

*Archive note: this article describes tools and policies as they stood on 17 January 2024. The GPT Store announcement below is translated from the Spanish source; its original English wording has not been verified. The English quotation in the postscript is reproduced as it appears in the original article.*



**Last week OpenAI**, the company behind ChatGPT, GPT-4 and DALL·E 3, among other models, **launched the “GPT Store”, a marketplace for customised GPT models**. For now, it is really more of a platform allowing paying ChatGPT users to discover and use customised versions of ChatGPT: monetisation has not yet been enabled, and its rollout will begin in the United States.



[As the launch article explains](https://openai.com/blog/introducing-the-gpt-store):



> “In just two months since GPTs were announced, users and builders have created more than 3 million versions, many of which have been shared for public use. The store features a wide range of GPTs developed by partners and the community, in categories such as DALL·E, writing, research, programming, education and lifestyle. Useful, impactful GPTs will be highlighted each week, and builders will be able to include their own GPTs in the store by following certain steps and complying with usage policies and GPT brand guidelines.”



**Given this significant development**—which many describe as *AI’s Apple Store or Play Store moment*, which Microsoft has already partly copied for its new Copilot Pro, and which Google has also announced it will introduce soon—**we decided to explore the enormous range of GPT proto-agents to find potentially interesting examples specifically related to politics and/or government**. Our interest lies less in their performance itself—we have not yet had time to test it thoroughly, and it will change considerably as the underlying AI model improves, something OpenAI repeatedly stresses to entrepreneurs in this field—than **in the needs they seek to address and the areas in which they operate**. We also hope **they will give those of you working in these sectors ideas for similar developments adapted to your countries, regions or professional expertise**.



## 1. GPT agents for electoral law



We have found several GPTs that work directly with their countries’ electoral legislation, including [Pakistan](https://chat.openai.com/g/g-Mjfj3mj0Y-electogpt-by-code-for-pakistan), [Brazil](https://chat.openai.com/g/g-Q3hzRfgtD-legis-expert-eleitoral) and [Argentina](https://chat.openai.com/g/g-xScDdElYC-asistente-electoral). The last of these, shown below, explicitly states that it focuses on common situations:



[![The Argentine Asistente electoral GPT, with four suggested questions about voting procedures.](https://elcontemplador.github.io/estrategia-english/assets/images/8479a2a4-edf6-44b4-a067-6bd454f82a09_895x624.png)](https://substackcdn.com/image/fetch/$s_!2qaf!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8479a2a4-edf6-44b4-a067-6bd454f82a09_895x624.png)



*Screenshot translation: “Electoral assistant. Electoral assistant in Argentina focusing on common situations.” By HERNAN A GAVIDIA L. Suggested questions: “What should be done if a voter shows their vote?”; “What is the procedure if a voter refuses to vote?”; “How should suspected fraudulent voting be handled?”; “What should be done if counterfeit ballot papers are detected?” The interface warns: “ChatGPT can make mistakes. Consider checking important information.”*



[ChatGPT — Asistente electoral](https://chat.openai.com/g/g-xScDdElYC-asistente-electoral)



## 2. Legislative assistants for citizens



We have also found several interesting examples in this field, some more general, such as one covering the laws of the state of [Ohio](https://chat.openai.com/g/g-HrfbNmTMK-state-law-impact), and others more specific, such as this assistant focusing on the social benefits a person may be eligible for in [Chile](https://chat.openai.com/g/g-G6dqcopjv-beneficios-sociales-chile):



[![Beneficios sociales Chile, a GPT offering guidance on Chilean social benefits.](https://elcontemplador.github.io/estrategia-english/assets/images/5a5bd72d-9007-4a94-8bc8-5f0ef8ba5f7c_925x631.png)](https://substackcdn.com/image/fetch/$s_!Geen!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5a5bd72d-9007-4a94-8bc8-5f0ef8ba5f7c_925x631.png)



*Screenshot translation: “Social benefits in Chile. Personalised adviser and guide to social benefits in Chile.” By jorgenova.cl. Suggested questions: “What types of grants and subsidies does the government offer?”; “Subsidy for home improvements”; “How do I access unemployment insurance?”; “How do I apply for Explicit Health Guarantees?” The same warning about checking important information appears below.*



[ChatGPT — Beneficios sociales Chile](https://chat.openai.com/g/g-G6dqcopjv-beneficios-sociales-chile)



## 3. Assistants on political parties’ perspectives and proposals



We found this GPT agent particularly interesting. It presents the two sides of US politics, summarising Democratic and Republican perspectives on different issues quite accurately in the examples we tried.



[![Two Sides Politics presents contrasting perspectives and suggested political topics.](https://elcontemplador.github.io/estrategia-english/assets/images/d17d3d47-f770-4d94-a6b7-a0be2ac43096_918x619.png)](https://substackcdn.com/image/fetch/$s_!ClnB!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd17d3d47-f770-4d94-a6b7-a0be2ac43096_918x619.png)



*Screenshot transcription: “Two Sides Politics. Bridging divides by helping people see the ‘other’ side’s point of view.” By allais.org. Suggested topics: “Guns”; “Current crisis in Gaza (Oct 2023)”; “Election fraud”; “Joe Biden”. The interface also warns that ChatGPT can make mistakes.*



[ChatGPT — Two Sides Politics](https://chat.openai.com/g/g-hiGqF87OL-two-sides-politics)



This kind of assistant, or something very similar, could also be built for election manifestos using comparable instructions. It would be very interesting as a conversational way for any prospective voter to compare the views of different parties or candidates.



## 4. Assistants for analysing election results



We also came across an agent whose proposition is more interesting than its current actual performance. It aims to offer international electoral and political analysis.



[![Political Panorama, a GPT presenting itself as an expert on international elections.](https://elcontemplador.github.io/estrategia-english/assets/images/445b0652-4435-4352-b0bf-00a1dcc4429e_903x637.png)](https://substackcdn.com/image/fetch/$s_!yl-A!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F445b0652-4435-4352-b0bf-00a1dcc4429e_903x637.png)



*Screenshot transcription: “Political Panorama. Expert on global elections, offering unbiased analysis and real-time updates.” By Brave Studio. Suggested prompts include “Tell me about the latest election results in France” and “Give me an update on the political situation in Japan”. Two others are cut off in the original image: “Can you provide an analysis of the recent election in …” and “What are the main strategies used in the US presiden…”. The interface warns that ChatGPT can make mistakes.*



[ChatGPT — Political Panorama](https://chat.openai.com/g/g-1aXBostFA-political-panorama)



## 5. Assistants acting as consultants or advisers



Finally, we found some early GPTs created by specific individuals particularly intriguing. They offer GPTs supposedly based on their professional knowledge—although we do not know whether or how that knowledge has been incorporated into their programming. Examples include an [electoral lawyer in Ecuador](https://chat.openai.com/g/g-XwCAOQbOH-abogado-electoral-ecuatoriano) and this proto-agent, “Frank”, an expert in election campaigns:



[![Frank, a GPT described as an expert in electoral campaign strategies.](https://elcontemplador.github.io/estrategia-english/assets/images/1f363a53-79de-4eec-80cc-8c98f8f6c61b_903x617.png)](https://substackcdn.com/image/fetch/$s_!u4b_!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1f363a53-79de-4eec-80cc-8c98f8f6c61b_903x617.png)



*Screenshot transcription: “Frank — Expert in Electoral Campaign Strategies.” By KAROLYNE P M FERRO. Suggested prompts: “How to integrate social media in electoral campaigns?”; “Can you analyze this campaign case study?”; “What are key factors in successful campaign strategi…” (cut off in the original); “How to navigate electoral legislation in campaigns?” The interface warns that ChatGPT can make mistakes. These are historical examples of the tool’s advertised prompts.*



[ChatGPT — Frank — Expert in Electoral Campaign Strategies](https://chat.openai.com/g/g-mjuHbnyem-frank-expert-in-electoral-campaign-strategies)



## Bonus proto-agent



Besides these five particularly interesting categories of GPTs, **we cannot miss the opportunity to recommend version 2.0 of Orador Inspirado**, our speech-drafting agent, **which we have made fully public and added to OpenAI’s GPT Store**.



[![Orador Inspirado V 2.0, Fernando Nieto’s speech-drafting GPT.](https://elcontemplador.github.io/estrategia-english/assets/images/45601b07-3db8-4f6b-800a-765b066dc229_996x684.png)](https://substackcdn.com/image/fetch/$s_!3T-4!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F45601b07-3db8-4f6b-800a-765b066dc229_996x684.png)



*Screenshot translation: “Orador Inspirado V 2.0. Specialist in creating first drafts for high-quality speeches. Press ‘Activate Orador Inspirado’ to begin.” By Fernando Nieto. The suggested opening prompt reads “ACTIVATE ORADOR INSPIRADO”.*



[ChatGPT — Orador Inspirado V 2.0](https://chat.openai.com/g/g-3TQXKuTeE-orador-inspirado-v-2-0)



**P.S.** Just a few hours before we closed this newsletter, [OpenAI published its policies for the 2024 elections](https://openai.com/blog/how-openai-is-approaching-2024-worldwide-elections). Alongside measures aimed at tackling disinformation and deepfakes, it announced something that may directly affect certain kinds of political GPTs—temporarily, we hope:



> We’re still working to understand how effective our tools might be for personalized persuasion. Until we know more, we don’t allow people to build applications for political campaigning and lobbying.



Fernando Nieto Lobato



*Director of Digital Innovation at Institución Educativa ALEPH*