# Perplexity — How to use AI for studying: 5 methods that work

- Company: Perplexity (perplexity.ai)
- Announced: 2026-10-15
- Category: not stated
- Coverage: not counted
- Announcement: no
- Group: routine
- Source: https://www.perplexity.ai/hub/blog/how-to-use-ai-for-studying
- Record: https://forck.live/items/26980-how-to-use-ai-for-studying-5-methods-that-work
- Subject: Perplexity
- Models affected: GPT-4

Perplexity's guide sets out five research-backed ways students can use AI to study, arguing that how a chatbot is used matters more than the tool itself. The methods are retrieval practice, distributed practice, self-explanation, interleaving and metacognitive review, each paired with a workflow, copy-ready prompts and an independent check done with the chat closed. Perplexity cites a 2025 randomized field experiment in PNAS (Bastani et al.) with roughly 1,000 high-school math students across four practice sessions: grades rose 48% with standard GPT-4 access and 127% with a guardrailed tutor that gave hints instead of full solutions. When access was removed, the standard-access group scored 17% lower than students who had never used AI at all, which Perplexity attributes to outsourced reasoning rather than practiced reasoning. A separate review pooling 35 studies and 4,193 participants found students who studied with AI learned moderately more on average, with results varying by subject, duration and manner of use. Perplexity notes the PNAS percentages come from high-school math and should not be generalized to other subjects or age groups.

## Evidence

Verbatim from https://www.perplexity.ai/hub/blog/how-to-use-ai-for-studying:

> Standard (plain and unmodified) GPT-4 access
> 
> Grades went up 48%
> 
> Guardrailed tutor (hints, not full solutions)
> 
> Grades went up 127%
> 
> Access removed after the practice sessions ended
> 
> Standard-access group scored 17% lower than students who had never used AI at all

---

Record: https://forck.live/items/26980-how-to-use-ai-for-studying-5-methods-that-work
Catalogue: https://forck.live/llms.txt
Current issue: https://forck.live/feed.md