[Submitted on 6 Apr 2026 (
[v1](https://arxiv.org/abs/2604.04721v1)), last revised 13 Oct 2026 (this version, v6)]
[View PDF](https://arxiv.org/pdf/2604.04721)
[HTML (experimental)](https://arxiv.org/html/2604.04721v6)
Abstract:People often optimize for long-term goals in collaboration: A mentor or companion doesn't just answer questions, but also scaffolds learning, tracks progress, and prioritizes the other person's growth over immediate results. In contrast, current AI systems are fundamentally short-sighted collaborators -- optimized for providing instant and complete responses, without ever saying no (unless for safety reasons). What are the consequences of this dynamic? Here, through a series of randomized controlled trials on human-AI interactions (N=1,222), we provide causal evidence for two key consequences of AI assistance: reduced persistence and impairment of unassisted performance. Across a variety of tasks, including mathematical reasoning and reading comprehension, we find that although AI assistance improves performance in the short-term, people perform significantly worse without AI and are more likely to give up. These findings are particularly concerning because persistence is foundational to skill acquisition and is one of the strongest predictors of long-term learning. We posit that persistence is reduced because AI conditions people to expect immediate answers, thereby denying them the experience of working through challenges on their own. These results add to the growing literature on cognitive off and AI-induced deskilling, and suggest the need for AI model development to prioritize scaffolding long-term competence alongside immediate task completion.
Submission history #
From: Grace Liu [
[view email](https://arxiv.org/show-email/ddcff5a0/2604.04721)]
**Mon, 6 Apr 2026 14:43:48 UTC (3,830 KB)**
[\[v1\]](https://arxiv.org/abs/2604.04721v1)
**Tue, 7 Apr 2026 03:41:15 UTC (3,830 KB)**
[\[v2\]](https://arxiv.org/abs/2604.04721v2)
**Tue, 4 Aug 2026 17:21:25 UTC (3,841 KB)**
[\[v3\]](https://arxiv.org/abs/2604.04721v3)
**Wed, 5 Aug 2026 10:45:02 UTC (3,841 KB)**
[\[v4\]](https://arxiv.org/abs/2604.04721v4)
**Sat, 3 Oct 2026 17:03:35 UTC (3,841 KB)**
[\[v5\]](https://arxiv.org/abs/2604.04721v5)
**[v6]** Tue, 13 Oct 2026 19:31:44 UTC (3,842 KB)
Additional Features
References & Citations
...
Bibliographic Explorer
(What is the Explorer?) Connected Papers
(What is Connected Papers?) Litmaps
(What is Litmaps?) scite Smart Citations
(What are Smart Citations?) alphaXiv
(What is alphaXiv?) CatalyzeX Code Finder for Papers
(What is CatalyzeX?) DagsHub
(What is DagsHub?) Gotit.pub
(What is GotitPub?) Hugging Face
(What is Huggingface?) ScienceCast
(What is ScienceCast?) Influence Flower
(What are Influence Flowers?) CORE Recommender
(What is CORE?) arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.