धड़धड़ाहट सुनाई दे रही है… साइंस की दुनिया में! वो दौर जब AI ने research papers को ‘slop’ बना दिया था—झूठे citations, बिना edit किये हुए AI ke jawab, aur bekaar ke diagrams… सब कुछ peer review ko chakma de raha tha. Par ab arXiv ne kaha hai—‘Bass karo! AI se bana hua junk submit karoge toh ek saal ke liye arXiv ke darwaze band, aur har baar peer review ke baad hi publish hoga tumhara paper.’
Kalpanā kījiye: ek aisa academic server jahan har roz hazaar papers chalte hain, lekin ab AI ke ‘garbage’ ne uski shaan ko khokhla kar rakh diya hai. Thomas Dietterich—Oregon State University ke emeritus professor aur arXiv ke editorial advisory council ke member—ne X (pehle Twitter) par ek thread choda, jisme unhone kaha, ‘AI se generate content submit karne walon ko ek saal ka break mil jayega… aur baad mein bhi har baar peer review se guzarna hoga.’
Hairān karne wali baat ye hai ki AI ke ‘slop’ ne sirf journals ko hi nahi, balki preprint servers ko bhi chakma de rakha hai. ArXiv ke moderation team ne ye policy isliye banayi hai kyunki AI ke jawab toh ‘nonsensical diagrams’ aur ‘fake citations’ ke saath aate hain—jo ki science ki integrity ko hi khokhla kar rahe hain. Dietterich ne apne thread mein likha, ‘Submissions ko scholarly communication ke standards par khara utarna hoga. AI ke bina edit kiye jawab, references ki galatiyan, aur bekaar ke figures—sabko reject kar diya jayega.’
Par sawaal ye hai: kya ye policy kafi hai? Kya AI ke ‘slop’ ko roka ja sakega? Ya phir science ki duniya ab AI ke ‘garbage’ ko hi apna leegi?
**Dastān shuru hoti hai yahan se…**
ArXiv, jo ki physics aur astronomy ka sabse bada preprint server hai, ne AI-generated content par kathor rules laga diye hain. Dietterich ne kaha, ‘AI se bana hua content submit karne walon ko ek saal ke liye arXiv se door rakha jayega. Aur uske baad bhi, har paper ko peer review se guzarna hoga—tabhi wo publish hoga.’
Ye policy isliye banayi gayi hai kyunki AI ke ‘slop’ ne science ki duniya mein ek ‘khauf’ paida kar diya hai. Fake citations, unedited AI responses, aur nonsensical diagrams—sab kuch peer review ko chakma de raha tha. ArXiv ke moderators ko lagta hai ki AI ke bina human touch ke papers science ki credibility ko khokhla kar rahe hain.
Par kya ye policy kafi hai? Kya AI ke ‘slop’ ko roka ja sakega? Ya phir science ki duniya ab AI ke ‘garbage’ ko hi apna leegi?
**History ki baat karein toh…**
AI ke ‘slop’ ka mahaul science ki duniya mein 2023 se hi tezi se badh raha hai. ChatGPT aur aur bhi AI tools ke aane ke baad, researchers ne AI se papers generate karne shuru kar diye—lekin unmein ‘garbage’ ki maatra itni zyada thi ki peer reviewers ko pata hi nahi chalta tha ki kya reject karna hai aur kya accept. 2024 tak, AI-generated papers ne journals ko chakma de diya tha—kuch toh ‘fake citations’ ke saath aate the, kuch toh ‘nonsensical diagrams’ ke saath.
ArXiv ne isko roka hai—lekin kya ye kafi hoga? Ya phir science ki duniya ab AI ke ‘garbage’ ko hi apna leegi?
**Aapki soch kya hai?**
AI ke ‘slop’ ko roka ja sakega? Ya phir science ki duniya ab AI ke ‘garbage’ ko hi apna leegi?
**Aur ek sawal…**
Kya aapko lagta hai ki AI se papers generate karne walon ko ek saal ka break milna chahiye? Ya phir ye policy zyada sakhit hai?
