Хиймэл оюун ухааны компаниуд өгөгдлийн хомсдолд орсноор ажиллагсдынхаа бүтээмжид тулгуурлаж байгаа ч тэдгээр ажилчид чанар муутай өгөгдөл нийлүүлэх болжээ.
Хиймэл оюун ухааны (AI) загваруудыг сургахад шаардлагатай цэвэр, анхдагч өгөгдлийн нөөц эрс багасч байна. Судалгаагаар AI сургалтад ашиглагдаж буй өгөгдлийн хэмжээ 2010 оноос хойш есөн сар тутамд хоёр дахин нэмэгдэж байгаа ч энэхүү өсөлт нь хязгаартаа тулж байна. Үүнээс үүдэн технологийн компаниуд хямд гэрээт ажилчдыг хөлсөлж, тусгай даалгавар гүйцэтгүүлэх замаар шинэ өгөгдөл үүсгүүлэх болжээ.
Гэвч эдгээр гэрээт ажилчид ажлын ачаалал болон хөлс багатай нөхцөлөөс шалтгаалан өгөгдөл боловсруулах явцад өөр AI чатботуудыг ашиглах нь түгээмэл болсон байна. Ажилчдын ярьснаар, тэд алдаа гаргахаас зайлсхийх болон ажлаа хөнгөвчлөхийн тулд AI-ийн үүсгэсэн өгөгдлийг өөрийн бүтээл мэтээр компаниудад ирүүлдэг аж. “Алис” хэмээх нэрээр танилцуулсан ажилтан компаниуд энэ үйлдлийг хянахыг хичээдэг ч бүрэн зогсоох боломжгүй байгааг онцолжээ.
Шинжээчдийн үзэж буйгаар, AI-ийн тусламжтайгаар үүсгэсэн өгөгдлийг дахин AI сургалтад ашиглах нь “хиймэл оюун ухааны каннибализм” үүсгэж, LLM загваруудын тогтвортой байдлыг алдагдуулах аюултай. Технологийн салбарынхан бусдын контентыг зөвшөөрөлгүй ашиглан бүтээгдэхүүн хөгжүүлж ирсэн бол өдгөө өөрсдийн бүтээсэн тогтолцоондоо чанар муутай, хуурамч өгөгдөл нийлүүлэгдэх асуудалтай нүүр тулж байна.
Дэлгэрэнгүйг эх сурвалжаас харах
↓Эх сурвалжийг нээх ↓
For tech companies racing to be the king of the AI hill, there are few things more precious than raw, original data.
To keep the large language models underlying our favorite AI chatbots up to date, tech companies have to feed them reams of fresh inputs. As one study found, the amount of data being used to train AI has doubled every nine months since 2010 — exponential growth which may soon hit a wall as stores of clean data run critically low.
When there’s no more original content to pilfer, companies have started paying workers to generate fresh training data, offering them low-quality contracts to train AI in hyper-specific tasks like running weekly payroll for Broadway musicians. Others have been hired for to film themselves doing degrading or menial chores like folding laundry or distinctly adult activities.
Predictably, this growing workforce behind the AI boom has started cutting corners en masse, turning to other AI chatbots to supply the data meant to feed AI chatbots. Talking to New Scientist, numerous insiders said this practice of AI cannibalism — a method experts have long warned can destabilize LLMs — is shockingly commonplace.
“It’s very widespread,” a worker identified as Alice told NewSci. “Every company I’ve worked for has had explicit guidelines around it and they clearly do try to catch people out, so I think they do care. But I don’t think they can stop it.”
In other words, AI companies are learning an ironic lesson: after purloining everybody else’s content without permission to create a product that threatens employment across the economy, the new precariat they’ve created are using the same tech to do the few human tasks they still need in as lazy a fashion as possible.
Though workers have to be careful not to be too obvious, Alice says it isn’t hard to pass AI-generated data off as her own, provided she scrubs the obnoxious linguistic tics of chatbots like ChatGPT before she submits it. “It’s only the sloppiest of users that get caught,” the AI contractor told NewSci. “Anyone with a modicum of awareness around AI hallmarks can tell their output not to use them, and at that point what are you going to do?”
“If these companies want quality data, then they shouldoffer quality contracts,” Alice continued. “Instead they’re low-balling struggling people, employing them for the barest possible amount of time and tossing them aside as projects are finished with no warning.”
Other contractors told NewSci they use LLMs in order to avoid making mistakes and losing their gig entirely.
“I was terrified of not having an income source, and then after that, it just became easier to run everything through LLMs,” one explained. “For alot of the projects that I do now, it’s creating scenarios, so I will use one LLM to help me create the scenario and then I’ll use a different LLM to help me create the files that go along with the scenario. I do feel guilty but like I said, in the beginning it was more about trying to make sure I wasn’t making any errors.”
Whatever the reason, it’s clear workers aren’t above feeding AI companies a taste of their own slop —a situation which could have drastic consequences for the AI race as a whole.
More on AI: Cop Accused of Using AI to Fake Evidence
The post AI Companies Are Learning an Ironic Lesson as the People They Pay to Improve Their Chatbots Are Just Feeding AI Slop Into Them appeared first on Futurism.
