why is finding REAL sources harder than the actual assignment

the getsolved thing sounds clutch!
I built a mini ai research assistant for literature review on getsolved.ai — fed it my topic ("AI in education ethics 2024–2026") and told it "find only peer-reviewed, no blogs, prioritize open access".
It gave me a clean list of 10 solid sources without fake ones.
Then I export to Zotero. Way less frustration.
Anyone else customizing bots for research? How do you avoid it hallucinating journals?

Update: switched to Scite as my ai research assistant for students — it shows "smart citations" (supporting/contradicting/mentioning) so you see if a paper is actually trusted or debunked.
Super useful for lit review — I paste a claim and it pulls real context from 200M+ citations.
Free trial is decent, but paid for unlimited. Saved me from citing a controversial study last week.
Thoughts on Scite vs Elicit for credibility checks?

YouTube explainers are life, but for serious lit review I swear by ResearchRabbit.
It’s free ai research assistant that builds visual maps of papers — start with one good article, and it shows connected ones like a family tree.
Helps find hidden gems you’d never Google.
Exported my map to Zotero and had 30+ sources in 20 mins.
Anyone using it for thesis mapping? Does the personalization get better over time?

@marvel_and_math one more tip: if you're doing literature review, start with a good prompt in an ai research assistant for literature review like "summarize key findings from 2023–2026 papers on [topic], highlight contradictions and gaps, cite open access only".
Tools like Elicit or Paperguide eat that up and give structured output.
Saves so much time vs manual reading.
Anyone got horror stories of AI giving fake sources? How do you double-check?