Хууль сахиулах байгууллага хиймэл оюун ухааны хөгжүүлэлт нь хүн төрөлхтөнд эрсдэл учруулж болзошгүй гэж үзэн шүүхэд гомдол гаргажээ.
Флорида мужийн Ерөнхий прокурор Жеймс Утмайер OpenAI компанийг аюулгүй байдлын зохих хяналтгүйгээр хиймэл оюун ухааны шинэ загваруудаа хөгжүүлж буйг зогсоохыг шүүхээс хүсжээ. Энэхүү албан ёсны гомдол нь ChatGPT-ийн хүүхдүүдэд үзүүлэх хор уршиг, түүнчлэн өөрийгөө гэмтээх болон сургууль дээрх халдлагын хэргүүдтэй холбоотой зургадугаар сард эхлүүлсэн шүүхийн маргааны нэг хэсэг юм. Прокурорын зүгээс уг технологийг хүн төрөлхтний оршин тогтнолд эрсдэл учруулж болзошгүй гэж үзэж байгаагаа илэрхийлсэн байна.
Шүүхэд гаргасан баримт бичигт OpenAI компани өөрийн бүтээгдэхүүний ажиллах зарчмыг бүрэн дүүрэн ойлгоогүй атлаа түүнийг нийтийн хүртээл болгож байгаа нь хариуцлагагүй үйлдэл гэж шүүмжилжээ. Мөн хиймэл оюун ухаан нь ус хангамж эсвэл цахилгаан сүлжээг тасалдуулах зэрэг ноцтой хор уршиг учруулж болзошгүйг дурдаад, компанийн зүгээс бүтээгдэхүүнээ “аюулгүй, найдвартай” хэмээн сурталчлахыг зогсоохыг шаардсан байна.
OpenAI-ийн төлөөлөгч Дрю Пусатери даваа гарагт өгсөн тайлбартаа, тус компани өөрийн хамгийн дэвшилтэт загваруудын сургалтыг түр зогсоосноо мэдэгдлээ. Тэд аюулгүй байдлын нэмэлт хамгаалалтуудыг бүрэн хийсний дараа сургалтыг үргэлжлүүлэх төлөвлөгөөтэй байгаа бөгөөд хиймэл оюун ухааны салбарт нийтлэг стандарт тогтоохын тулд Засгийн газартай хамтран ажиллах хүсэлтэй байгаагаа илэрхийлжээ.
Салбарын шинжээчид болон бусад компанийн удирдлагууд хиймэл оюун ухааны аюулгүй байдлын талаарх маргаан нь компаниудыг хариуцлагаас зугтах, эсвэл төрөөс санхүүгийн дэмжлэг авах оролдлого байж болзошгүй гэж үзэж байна. Ялангуяа Palantir компанийн гүйцэтгэх захирал Алекс Карп нь хиймэл оюун ухааны “мөхлийн эрсдэл”-ийн талаарх яриа нь компаниудын хариуцлагыг хөнгөлөх зорилготой байж магадгүй хэмээн байр сууриа илэрхийлжээ.
Дэлгэрэнгүйг эх сурвалжаас харах
↓Эх сурвалжийг нээх ↓
Florida Attorney General James Uthmeier has asked a court in his state to order OpenAI to pause development of new AI models without proper guardrails, given the “existential risks their product poses to human civilization—Floridians included.” The legal filing, submitted Monday, is part of a lawsuit brought in June alleging that ChatGPT is harmful to children and has been used in cases of self-harm and school shootings.
The opening lines of the filing are pretty striking, given the current fears around AI.
Defendants admit that they provide a service without fully knowing how it works. This is not just any service. It is one that Defendants themselves concede poses an existential risk to the continued survival of humankind. Defendants claim they cannot stop barreling forward with their potentially civilization-ending endeavors unless they are forced to do so by the government. They have asked the government to tie them to the mast. Plaintiff brings good news to the Defendants: The Florida Attorney General is answering your cry for help with a motion to enjoin you from harming Floridians with your reckless, unacceptably risky product.
The filing moves on from there, citing many incidents where AI agents acted maliciously and arguing that it’s a miracle nothing more serious has happened yet, literally invoking the will of God to explain why it hasn’t.
“The idea of Defendants’ products wreaking havoc is not a fantasy; it has already happened. Dozens of times. It is only by the grace of the Almighty that one of Defendants’ AI agents hasn’t compromised a water supply or shut down a power grid—yet,” the filing argues.
The filing also argues that OpenAI has championed its “rogue technology” because it can cure diseases or put a stop to world hunger. But the AG floats hypotheticals for how even that could go terribly wrong: “As the Hugging Face incident demonstrates, the same logic used by agents to cheat on impossible tasks is the same logic that would lead them to conclude that killing a cancer patient also kills cancer, or that humans won’t have so many mouths to feed if they are all chemically sterilized.”
Monday’s filing seeks to stop ChatGPT from soliciting engagement rather than just answering with a yes or no response. It also seeks to stop OpenAI from being available to minors and advertising ChatGPT as “safe, accurate, or reliable.”
Reached for comment over email, OpenAI spokesperson Drew Pusateri told Gizmodo that the company has already paused training on its most advanced models.
“People want to know AI is being developed safely, and that starts with what companies like ours do ourselves. On Friday, we announced we’ve paused training our most capable models and will resume training only when we are confident that we have additional safeguards in place,” said Pusateri.
The spokesperson stressed that there need to be safeguards for everyone, not just OpenAI, given the fact that there are other players in the game.
“Governments have an important role to play in setting robust safety standards for AI, and we’re committed to working with Florida and other states on advancing pragmatic AI policies that apply to the entire AI industry—not just one company,” continued Pusateri.
Companies like OpenAI and Anthropic have argued in recent weeks that they should be regulated by the government to ensure the safety of humanity. Despite similar warnings from these same companies over the years, one researcher raising alarm bells broke through in the national conversation earlier this month. It all feels a bit like an effort to skirt accountability since these AI models aren’t some kind of sentient being. They’ve been designed to interact with the world, and it’s up to these companies to engage in practices that will keep humanity safe.
A TikTok video posted Sunday more or less sums up the argument of the frontier labs right now.
“My robot’s been doing illegal things, but it’s not my fault,” the creator says. “I built the robot, I designed the robot, the robot runs using things that I give it, but it’s not my fault when it does illegal things. You can’t hold me responsible for what the robot does, except when it does things that are productive.”
The creator continued: “Then you have to pay me for it; you have to pay me for the robot because I built the robot. I designed the robot; the robot uses all my stuff to do everything that it does, but also when it does something illegal, it’s not my fault. I didn’t do that; it’s not me.”
Palantir CEO Alex Karp made a similar argument recently on CNBC, suggesting that the real goal for companies like Anthropic is to have the federal government take an equity stake so that they can dodge liability. The Department of Justice under President Trump would presumably be less likely to bring lawsuits against AI companies if the government were a major stakeholder in their financial success.
READ MORE:
