AI research is drowning in too many low-quality papers
AI conferences like NeurIPS are seeing a record flood of submissions—over 21,000 papers this year, more than double since 2020.
But with so many new papers, especially ones involving high schoolers through mentoring firms, people are starting to question if the quality is slipping.
Who's really writing these papers?
A recent grad says he supervised over 100 AI research papers this year, many co-authored by high school students.
This has sparked debate about whether these young co-authors really understand the work or if it's just padding resumes—raising concerns about what counts as real scientific contribution.
Why does it matter?
With so much research out there and less time for reviewers to check each paper carefully, it's getting harder for everyone—from scientists to regular readers—to tell which advances are legit.
All this noise makes it tough to spot real progress in AI and keep science trustworthy.