Lihat sumber Sora (model teks-ke-video)

'''Sora''' adalah sebuah [[model teks-ke-video]] yang dikembangkan oleh organisasi penelitian [[kecerdasan buatan]] (AI) [[OpenAI]]. Model ini menghasilkan video berdasarkan perintah deskriptif serta dapat memperluas video yang sudah ada.<ref>Cade Metz. [https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos]. ''The New York Times''. February 15, 2024.</ref><ref>Tim Brooks. [https://openai.com/research/video-generation-models-as-world-simulators Video generation models as world simulators]. ''OpenAI''. February 15, 2024.</ref> Sampai dengan Februari 2024, Sora masih belum tersedia luas untuk publik.<ref>Angela Yang. [https://www.nbcnews.com/tech/tech-news/openai-sora-video-artificial-intelligence-unveiled-rcna139065 OpenAI teases 'Sora,' its new text-to-video AI model]. ''NBC News''. February 15, 2024.</ref>

== Sejarah ==
Beberapa model teks-ke-video lainnya telah dibuat sebelum Sora, diantaranya Make-A-Video dari [[Meta Platforms|Meta]], Gen-2 dari Runway, dan Lumiere dari [[Google]], yang sampai pada bulan Februari 2024, juga masih dalam tahap penelitian.<ref>Shankar Das. [https://epiexpress.com/sora-ai-can-generate-realistic-video-based-on-text-prompts-openai/ Sora AI Can Generate Realistic Video based on Text Prompts: OpenAI]. ''Epi Express''. February 15, 2024.</ref><ref>[https://www.wired.com/story/openai-sora-generative-ai-video sumber].</ref> [[OpenAI]], sebelum mengumumkan Sora, telah merilis DALL·E 3, sebuah [[model teks-ke-gambar]] pada bulan September 2023.<ref>Lisa Lacy. [https://www.cnet.com/tech/meet-sora-openais-text-to-video-generator/ Meet Sora, OpenAI's Text-to-Video Generator]. ''CNET''. February 15, 2024.</ref>

Tim yang mengembangkan Sora menamakannya setelah [[wiktionary:空#Japanese|kata dalam bahasa Jepang untuk langit]] sebagai penanda untuk "potensi kreatif yang tak terbatas".<ref>Cade Metz. [https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos]. ''The New York Times''. February 15, 2024. Metz, Cade (February 15, 2024). [https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html "OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos"]. ''The New York Times''. [https://web.archive.org/web/20240215220626/https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html Archived] from the original on February 15, 2024 . Retrieved February 15, 2024 .</ref> Pada tanggal 15 Februari 2024, OpenAI pertama kali memperlihatkan Sora dengan merilis beberapa klip [[video definisi tinggi]] yang dibuatnya, diantaranya termasuk [[Mobil SUV|SUV]] yang melaju di jalan pegunungan, animasi "monster berbulu pendek" di samping lilin, dua orang berjalan melewati [[Tokyo]] di tengah salju, dan rekaman rekaan [[Demam Emas California|demam emas California]], dan menyatakan bahwa Sora mampu menghasilkan video berdurasi hingga satu menit.<ref>Levy, Steven (February 15, 2024). [https://www.wired.com/story/openai-sora-generative-ai-video/ "OpenAI's Sora Turns AI Prompts Into Photorealistic Videos"]. ''Wired''. [https://web.archive.org/web/20240215234655/https://www.wired.com/story/openai-sora-generative-ai-video/ Archived] from the original on February 15, 2024 . Retrieved February 16, 2024 .</ref><ref>Shankar Das. [https://epiexpress.com/sora-ai-can-generate-realistic-video-based-on-text-prompts-openai/ Sora AI Can Generate Realistic Video based on Text Prompts: OpenAI]. ''Epi Express''. February 15, 2024. Shankar Das (February 15, 2024). [https://epiexpress.com/sora-ai-can-generate-realistic-video-based-on-text-prompts-openai/ "Sora AI Can Generate Realistic Video based on Text Prompts: OpenAI"]. ''Epi Express''. [https://web.archive.org/web/20240215234018/https://mashable.com/article/openai-sora-ai-text-to-video-model-announcement Archived] from the original on February 15, 2024 . Retrieved February 16, 2024 .</ref> Laporan teknis kemudian disebarkan, yang menyoroti metode yang digunakan untuk melatih model tersebut.<ref>Tim Brooks. [https://openai.com/research/video-generation-models-as-world-simulators Video generation models as world simulators]. ''OpenAI''. February 15, 2024. Brooks, Tim; Peebles, Bill; Holmes, Connor; DePue, Will; Guo, Yufei; Jing, Li; Schnurr, David; Taylor, Joe; Luhman, Troy; Luhman, Eric; Ng, Clarence Wing Yin; Wang, Ricky; Ramesh, Aditya (February 15, 2024). [https://openai.com/research/video-generation-models-as-world-simulators "Video generation models as world simulators"]. ''OpenAI''. [https://web.archive.org/web/20240216072133/https://openai.com/research/video-generation-models-as-world-simulators Archived] from the original on February 16, 2024 . Retrieved February 16, 2024 .</ref><ref>Benj Edwards. [https://arstechnica.com/information-technology/2024/02/openai-collapses-media-reality-with-sora-a-photorealistic-ai-video-generator/ OpenAI collapses media reality with Sora, a photorealistic AI video generator]. ''Ars Technica''. February 16, 2024.</ref> CEO OpenAI [[Sam Altman]] juga memposting serangkaian tweet, menanggapi permintaan pengguna [[Twitter]] dengan video yang dihasilkan oleh Sora.

OpenAI telah menyatakan bahwa mereka berencana untuk membuat Sora tersedia untuk umum tetapi tidak disebutkan kapan akan tersedia.<ref>Levy, Steven (February 15, 2024). [https://www.wired.com/story/openai-sora-generative-ai-video/ "OpenAI's Sora Turns AI Prompts Into Photorealistic Videos"]. ''Wired''. [https://web.archive.org/web/20240215234655/https://www.wired.com/story/openai-sora-generative-ai-video/ Archived] from the original on February 15, 2024 . Retrieved February 16, 2024 .</ref><ref>Angela Yang. [https://www.nbcnews.com/tech/tech-news/openai-sora-video-artificial-intelligence-unveiled-rcna139065 OpenAI teases 'Sora,' its new text-to-video AI model]. ''NBC News''. February 15, 2024. Yang, Angela (February 15, 2024). [https://www.nbcnews.com/tech/tech-news/openai-sora-video-artificial-intelligence-unveiled-rcna139065 "OpenAI teases 'Sora,' its new text-to-video AI model"]. ''NBC News''. [https://web.archive.org/web/20240215235542/https://www.nbcnews.com/tech/tech-news/openai-sora-video-artificial-intelligence-unveiled-rcna139065 Archived] from the original on February 15, 2024 . Retrieved February 16, 2024 .</ref> Perusahaan memberikan akses terbatas kepada sebuah "tim merah" kecil, termasuk para ahli di bidang misinformasi dan bias, untuk melakukan pengujian terhadap model tersebut.<ref>Lisa Lacy. [https://www.cnet.com/tech/meet-sora-openais-text-to-video-generator/ Meet Sora, OpenAI's Text-to-Video Generator]. ''CNET''. February 15, 2024. Lacy, Lisa (February 15, 2024). [https://www.cnet.com/tech/meet-sora-openais-text-to-video-generator/ "Meet Sora, OpenAI's Text-to-Video Generator"]. ''CNET''. [https://web.archive.org/web/20240216004932/https://www.cnet.com/tech/meet-sora-openais-text-to-video-generator/ Archived] from the original on February 16, 2024 . Retrieved February 16, 2024 .</ref> Perusahaan juga berbagi Sora dengan sekelompok kecil profesional kreatif, termasuk pembuat video dan seniman, untuk mencari masukan mengenai kegunaannya di bidang kreatif.<ref>Will Douglas Heaven. [https://www.technologyreview.com/2024/02/15/1088401/openai-amazing-new-generative-ai-video-model-sora/ OpenAI teases an amazing new generative video model called Sora]. ''MIT Technology Review''. February 15, 2024.</ref>

== Kemampuan dan keterbatasan ==
Teknologi di balik Sora merupakan pengembangan dari teknologi di balik DALL-E 3. Menurut OpenAI, Sora adalah transformator difusi<ref>William Peebles. [https://openaccess.thecvf.com/content/ICCV2023/html/Peebles_Scalable_Diffusion_Models_with_Transformers_ICCV_2023_paper.html 2023 IEEE/CVF International Conference on Computer Vision (ICCV)]. 2023. hlm. 4172–4182. doi:10.1109/ICCV51070.2023.00387. ISBN 979-8-3503-0718-4.</ref> – model difusi laten denoising dengan satu transformer sebagai denoiser. Sebuah video dihasilkan dalam ruang laten dengan "tambalan" 3D, kemudian diubah menjadi ruang standar oleh dekompresor video. Pembuatan teks ulang digunakan untuk menambah data pelatihan, dengan menggunakan model video-ke-teks untuk membuat deskripsi mendetail pada video.<ref>Benj Edwards. [https://arstechnica.com/information-technology/2024/02/openai-collapses-media-reality-with-sora-a-photorealistic-ai-video-generator/ OpenAI collapses media reality with Sora, a photorealistic AI video generator]. ''Ars Technica''. February 16, 2024. Edwards, Benj (February 16, 2024). [https://arstechnica.com/information-technology/2024/02/openai-collapses-media-reality-with-sora-a-photorealistic-ai-video-generator/ "OpenAI collapses media reality with Sora, a photorealistic AI video generator"]. ''Ars Technica''. [https://web.archive.org/web/20240217000922/https://arstechnica.com/information-technology/2024/02/openai-collapses-media-reality-with-sora-a-photorealistic-ai-video-generator/ Archived] from the original on February 17, 2024 . Retrieved February 17, 2024 .</ref>

OpenAI melatih model tersebut menggunakan video yang tersedia secara publik serta kumpulan video berhak cipta yang dilisensikan untuk tujuan tersebut, namun tidak mengungkapkan jumlah dan sumber pasti dari video tersebut.<ref>Cade Metz. [https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos]. ''The New York Times''. February 15, 2024. Metz, Cade (February 15, 2024). [https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html "OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos"]. ''The New York Times''. [https://web.archive.org/web/20240215220626/https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html Archived] from the original on February 15, 2024 . Retrieved February 15, 2024 .</ref> Setelah dirilis, OpenAI mengakui beberapa kekurangan Sora, termasuk kesulitannya dalam menyimulasikan fisika kompleks, memahami [[kausalitas]], dan membedakan antara kiri dari kanan.<ref>Antonio Pequeño IV. [https://www.forbes.com/sites/antoniopequenoiv/2024/02/15/openai-reveals-sora-ai-video-model-capable-of-realistic-text-to-video-prompts/ OpenAI Reveals 'Sora': AI Video Model Capable Of Realistic Text-To-Video Prompts]. ''Forbes''. February 15, 2024.</ref> Salah satu contoh menunjukkan sekelompok anak serigala yang tampak secara ganda, sehingga menciptakan skenario yang sulit diikuti.<ref>[https://ai-review.io/sora-review-free-ai-video-generator-from-openai/ Sora Review New AI Video Generator From OpenAI]. 2024-02-18.</ref> OpenAI juga menyatakan bahwa, sesuai dengan praktik keselamatan perusahaan yang ada, Sora akan membatasi teks yang berisi gambar seksual, kekerasan, kebencian, atau konten selebritas, serta konten yang menampilkan [[kekayaan intelektual]] yang sudah ada.<ref>Lisa Lacy. [https://www.cnet.com/tech/meet-sora-openais-text-to-video-generator/ Meet Sora, OpenAI's Text-to-Video Generator]. ''CNET''. February 15, 2024. Lacy, Lisa (February 15, 2024). [https://www.cnet.com/tech/meet-sora-openais-text-to-video-generator/ "Meet Sora, OpenAI's Text-to-Video Generator"]. ''CNET''. [https://web.archive.org/web/20240216004932/https://www.cnet.com/tech/meet-sora-openais-text-to-video-generator/ Archived] from the original on February 16, 2024 . Retrieved February 16, 2024 .</ref>

Tim Brooks, seorang peneliti Sora, menyatakan bahwa model tersebut menemukan cara membuat [[grafika komputer 3D]] hanya dari kumpulan datanya, sementara Bill Peebles, juga seorang peneliti Sora, mengatakan bahwa model tersebut secara otomatis membuat sudut video yang berbeda tanpa diminta sebelumnya.<ref>Levy, Steven (February 15, 2024). [https://www.wired.com/story/openai-sora-generative-ai-video/ "OpenAI's Sora Turns AI Prompts Into Photorealistic Videos"]. ''Wired''. [https://web.archive.org/web/20240215234655/https://www.wired.com/story/openai-sora-generative-ai-video/ Archived] from the original on February 15, 2024 . Retrieved February 16, 2024 .</ref> Menurut OpenAI, video yang dihasilkan oleh Sora ditandai dengan [[Inisiatif Keaslian Konten|metadata C2PA]] untuk menunjukkan bahwa video tersebut dihasilkan oleh kecerdasan buatan.<ref>Cade Metz. [https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos]. ''The New York Times''. February 15, 2024. Metz, Cade (February 15, 2024). [https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html "OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos"]. ''The New York Times''. [https://web.archive.org/web/20240215220626/https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html Archived] from the original on February 15, 2024 . Retrieved February 15, 2024 .</ref>

== Reaksi ==
Will Douglas Heaven dari ''MIT Technology Review'' menyebut video demonstrasi tersebut "mengesankan", namun menyatakan bahwa video tersebut pasti dipilih secara khusus dan mungkin tidak mewakili keluaran khas Sora pada umumnya.<ref>Will Douglas Heaven. [https://www.technologyreview.com/2024/02/15/1088401/openai-amazing-new-generative-ai-video-model-sora/ OpenAI teases an amazing new generative video model called Sora]. ''MIT Technology Review''. February 15, 2024. Heaven, Will Douglas (February 15, 2024). [https://www.technologyreview.com/2024/02/15/1088401/openai-amazing-new-generative-ai-video-model-sora/ "OpenAI teases an amazing new generative video model called Sora"]. ''MIT Technology Review''. [https://web.archive.org/web/20240215220619/https://www.technologyreview.com/2024/02/15/1088401/openai-amazing-new-generative-ai-video-model-sora/ Archived] from the original on February 15, 2024 . Retrieved February 15, 2024 .</ref> Akademisi Amerika Oren Etzioni menyatakan kekhawatirannya atas kemampuan teknologi tersebut dalam menciptakan [[disinformasi]] daring untuk kampanye politik.<ref>Cade Metz. [https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos]. ''The New York Times''. February 15, 2024. Metz, Cade (February 15, 2024). [https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html "OpenAI Unveils A.I. That Instantly Generates Eye-Popping Videos"]. ''The New York Times''. [https://web.archive.org/web/20240215220626/https://www.nytimes.com/2024/02/15/technology/openai-sora-videos.html Archived] from the original on February 15, 2024 . Retrieved February 15, 2024 .</ref> Untuk ''[[Wired (majalah)|Wired]]'', Steven Levy juga menulis bahwa film tersebut berpotensi menjadi "badai informasi yang salah" dan berpendapat bahwa klip pratinjaunya "mengesankan" tetapi "tidak sempurna" dan "menunjukkan pemahaman yang muncul tentang tata bahasa sinematik" karena perubahan tangkapan kamera tanpa diminta. Levy menambahkan, "akan memakan waktu yang sangat lama, jika akan terjadi, sebelum teks-ke-video mengancam pembuatan film yang sebenarnya."<ref>Levy, Steven (February 15, 2024). [https://www.wired.com/story/openai-sora-generative-ai-video/ "OpenAI's Sora Turns AI Prompts Into Photorealistic Videos"]. ''Wired''. [https://web.archive.org/web/20240215234655/https://www.wired.com/story/openai-sora-generative-ai-video/ Archived] from the original on February 15, 2024 . Retrieved February 16, 2024 .</ref> Lisa Lacy dari [[CNET]] menyebut bahwa contoh videonya "sangat realistis – kecuali mungkin saat wajah manusia terlihat dari dekat atau saat makhluk laut sedang berenang".<ref>Lisa Lacy. [https://www.cnet.com/tech/meet-sora-openais-text-to-video-generator/ Meet Sora, OpenAI's Text-to-Video Generator]. ''CNET''. February 15, 2024. Lacy, Lisa (February 15, 2024). [https://www.cnet.com/tech/meet-sora-openais-text-to-video-generator/ "Meet Sora, OpenAI's Text-to-Video Generator"]. ''CNET''. [https://web.archive.org/web/20240216004932/https://www.cnet.com/tech/meet-sora-openais-text-to-video-generator/ Archived] from the original on February 16, 2024 . Retrieved February 16, 2024 .</ref>

== Lihat pula ==
* [[ChatGPT]]

== Pranala luar ==
*

== Referensi ==
<references />

== Sumber dan atribusi ==

Konten artikel ini diadaptasi dari [https://id.wikipedia.org/w/index.php?title=Sora+%28model+teks-ke-video%29&oldid=27946487 Wikipedia bahasa Indonesia], revisi 27946487 (2025-10-10T03:27:37Z), yang tersedia berdasarkan lisensi Creative Commons Atribusi-BerbagiSerupa (CC BY-SA). Mohon gunakan konten ini secara bijak serta sesuai dengan ketentuan lisensi yang berlaku.

<!-- WIKI_UNISSULA_PRESENTATION_V4 -->