SYSTEM NOTICE

Auto translation by AI. Be sure, accuracy, nuances and authorial intent may not be fully reflected.
見出し画像

Archives - 2025/09/27 30th Generative AI Report Meeting

CreativeEdge CL+

The September report meeting has concluded. Thank you for attending. The next meeting (31st) is scheduled for Saturday, October 25th, from 10:00 AM to 11:00 AM.


Download the materials from the day:

  • 20250927-GenAI-30.pdf (PDF / 173 pages / 13.2MB)

  • Download period: Until Friday, October 24, 2025, at 11:59 PM.
    The period has ended, and it has been moved and archived in Membership.

Slide materials


Archive video for general public (11 minutes 24 seconds)

  • This is a version of the live stream recording shortened to about 10 minutes. Anyone can watch it.
    *The verification video shown during the live stream has frame skipping.


Archive video for participants (63 minutes 37 seconds):

  • Please enter the viewing password provided on the day.

  • Viewing period: Until Friday, October 24, 2025, at 11:59 PM.
    The period has ended, and it has been moved and archived in Membership.


Past report meetings:


Future tasks:

  • Exploring ways to utilize Generative AI in the design/art field

  • Utilization of Generative AI in planning, design, and processes to enhance creativity

  • Verifying the potential for factors that diminish creators' motivation and examining what rules/guidelines are necessary

  • Held an opinion exchange meeting with advertising agencies and production companies regarding the issue of clients generating creative works such as photos and illustrations in-house (instead of commissioning professionals) due to the spread of generative AI

  • Regarding copyright/portrait rights, etc., and AI ethics

  • Regarding a mechanism for creators who do not want their work to be learned by AI to attach a 'Do Not Train' tag


Image Generation/Video Generation Policy (Updated June 2024 / Title changed: March 2024):

  1. Do not include author names or work titles in prompts (the same applies to film director names, movie titles, characters, actor names, etc.)

  2. Do not include the names of celebrities or brand names in prompts

  3. Do not use others' copyrighted works as references

  4. Do not use free materials provided on the internet as references

  5. Do not treat generated images as 'photographs' because hallucinations (e.g., errors in human joints or fingers, inaccuracies in light or shadows, etc.) occur

  6. Show respect for specific historically and culturally significant works and refrain from reproducing them (not due to copyright issues, but to respect the cultural value and influence the works possess)

  7. Do not publish illustration-style generated images, such as those from the Niji model, as your own work

  8. Do not use generated images as-is; use them as 'material for your work'

  9. When publishing, indicate that it was generated by AI


Excerpt from the day's materials:

It was predicted that photorealistic expression would be possible by 2016, but it is already achievable
Prompting of already created 3DCG characters
Image generation models and video generation models currently in use
Combination of reference images and prompts
Prompting of 3DCG characters
Video generation consumes many credits (it costs money)
Video expression through remixing multiple reference images
Video expression through remixing multiple reference images
Use 3DCG image references for areas where you want to maintain consistency (backgrounds)
Video with backgrounds expressed using 3DCG image references
Control of two characters
High-performance models (Google Veo 3, KLING) are required for video production
Video production of two characters
Components for video generation (remixing)
Results of remixing
Start by preparing the components of the scene
Generating ID cards by repeating the remix process four times
The concept of image processing has changed
We will release the latest technologies and features on our YouTube channel as soon as possible (starting next week)


Record:

Generative AI and Creative - 30th Online Report Meeting
~Creative Thinking for Creators Mastering AI~

  • Date and Time:Saturday, September 27, 2025, 10:00 AM - 11:00 AM

  • Online streaming (using Zoom)

  • Participation Fee:Free

  • Target Audience:Those interested in the impact of Generative AI on the creative industry

  • How to Participate:Go to the application page (registration closed)

  • Materials:The URL will be provided in the Zoom chat on the day of the event

  • Archive:There will be no public archive of the live stream, buta 10-15 minute digest version will be released at a later date. The full live archive is
    available to members.


Updated: Sunday, August 28, 2025 / Published: Saturday, September 27, 2025

いいなと思ったら応援しよう！