# AI Lacks Independent Learning, No Existential Threat, New Study Finds 🤖

**URL:** <https://onehack.st/t/ai-lacks-independent-learning-no-existential-threat-new-study-finds/289036>\
**Category:** News & Articles\
**Tags:** ai\
**Created:** [August 15, 2024, 3:28am UTC](https://onehack.st/t/ai-lacks-independent-learning-no-existential-threat-new-study-finds/289036 "2024-08-15T03:28:20Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![TheJoker](https://onehack.st/user_avatar/onehack.st/thejoker/32/22660_2.png) [@TheJoker](https://onehack.st/u/TheJoker)\
**Post date:** [August 15, 2024, 3:28am UTC](https://onehack.st/t/ai-lacks-independent-learning-no-existential-threat-new-study-finds/289036/1 "2024-08-15T03:28:20Z")

</div>

**Summary:**

* * *

1. **Limited Learning Abilities**  
New research reveals that large language models (LLMs), like ChatGPT, cannot independently learn or acquire new skills without explicit instructions, debunking fears of them evolving complex reasoning abilities.

2. **Predictable and Controllable**  
The study shows that LLMs remain controllable and predictable, with no evidence supporting the emergence of complex thinking skills that could pose existential threats.

3. **Focus on Misuse Risks**  
While the study clears up concerns about LLMs posing existential threats, it highlights the need to address potential misuse, such as generating fake news, as a significant area for future research. 🛡

**Read more at:** [Neuroscience News](https://neurosciencenews.com)