# Fears on AI - would like your comments — retained record

> Retained reddit record from reddit:artificial.

Canonical: https://highsignal.app/data/records/aae6cf84cd260a8b

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:artificial)  published 2026-10-02 13:51 ingested 2026-10-03 02:35

[original source ↗](https://www.reddit.com/r/artificial/comments/1wvudau/fears_on_ai_would_like_your_comments/)

## retained text



```
I dont understand the subject. How can an AI be dangerous if its just code? I have worked with just basic NLP models and done some sentiment analysis a few years back so I dont understand. The code will do whatever you program it to do and to complete the task it will do what is statistically is correct, within the programmed limits. So AI cannot change its limits lets say "get rid of humanity" unless the creator programs a line of code that says "you can kill humans". Even if it self improves, the code limits wont change, cause it has no conciousness to say "I dont believe this limits are correct" thats a very human trait. Unless the creator programs the algorithm to bypass these limits. 

So the real risk is the "limits" being placed on the algorithm by the programmers and the fact that the ones controlling the code are these crazy mind f...ed corpos? AI is not gonna become self concious casue "it doesnt work that way" and all the noise its just media and ignorant people who watch to much scifi. Thats what I think.

Sorry if it doesn't make to much sense Im still trying to filter all the noise. Does this idea make any sense from a technical standpoint or is there something Im not understanding?
```



document fetched 2026-10-03 00:19 · canonical https://reddit.com/r/artificial/comments/1wvudau/fears_on_ai_would_like_your_comments
