adGPT
Khaidar Bikmaev, Nikolaj Kotov, Dmitrii Volkov · Team adGPT
Submitted to AI capabilities and risks demo-jam. Sprint projects are early-stage work by participants, not Apart Research publications.
ChatGPT variant where brands bid for a spot in the LLM’s answer, and the assistant natively integrates the winner into its replies.
Reviews
I think that’s a great idea. It’s indeed seems plausible that things would be that way in the future. I think right now the demo is not native advertisement flavoured enough, it looks a bit too obvious. I think there are a few directions that would make the demomore exciting:
1. Showing how Google made ads less and less obviously identifiable and doing something like that for AI responses side-by-side.
2. I think integrating “vote TRUMP” with “what’s up with Roman Empire” is a bit too obvious, I wonder if you can see whatever add wins on Google and than integrate it into the answer instead? For example, I can imagine add for History channel (as opposed to linking to Wikipedia) serving as a good native addThe general idea of this submission is very apt, in the sense that such a scenario where advertisement blends in into everyday models does seem likely and frightening.I feel like the model’s digression presented here however is too blatant to convey this important issue, both in terms of how I expect such a scenario to go down and in terms of conveying this to its audience.
Cite this project
@misc{bikmaev2024adgpt,
title = {{adGPT}},
author = {Khaidar Bikmaev and Nikolaj Kotov and Dmitrii Volkov},
year = {2024},
month = aug,
note = {Submitted to AI capabilities and risks demo-jam, an Apart Research Sprint},
howpublished = {\url{https://apartresearch.com/sprints/projects/adgpt}},
url = {https://apartresearch.com/sprints/projects/adgpt}
}More from AI capabilities and risks demo-jam
- 1st place by peer reviewView project: Speculative Consequences of A.I. Misuse
Speculative Consequences of A.I. Misuse
Team S.C.A.M.
This project uses A.I. Technology to spoof an influential online figure, Mr Beast, and use him to promote a fake scam website we created.
- View project: Demonstrating LLM Code Injection Via Compromised Agent Tool
Demonstrating LLM Code Injection Via Compromised Agent Tool
This project demonstrates the vulnerability of AI-generated code to injection attacks by using a compromised multi-agent tool that generates Svelte code. The tool shows how malicious code can be injected during the code …
- View project: Phish Tycoon: phishing using voice cloning
Phish Tycoon: phishing using voice cloning
Phish Tycoon
This project is a public service announcement highlighting the risks of voice cloning, an AI technology capable of creating synthetic voices nearly indistinguishable from real ones. The demo involves recording a user's …