AI being discussed at wider level across CERN: administration, etc
- facilitated by KT
- idea is to create an AI hub - a docking station for individual efforts
--
Need a statement on human oversight, responsibility and ownership
Need a statement/guidance on use of confidential data (personal, financial, sensitive) in AI prompts
--
Use cases discussion
-
- Image and video generation: photos, technical animations, artists’ impressions [Piotr]
- Design team - for prototyping
- Text to image tools for images (used in public events, science shows)
- Medialab - text to image tools; mostly not used for production
- Audio, image and video post-production [Piotr]
- Enhance images - in PremierPro
- Generate captions - integrated in PremierPro
- Text - results are always adapted/checked/verified by humans afterwards
- Research - to help understand technical concepts/historical context, for literature research - ChatGPT, scite.ai
- Inspiration - to help brainstorm titles etc - ChatGPT
- Editing - check and modify text e.g. to adapt to different age groups/target audiences/tone or shorten/expand - ChatGPT
- Proofing - Word, ChatGPT, Deepl, Grammarly
- Translations - Deepl, ChatGPT
- Generation - produce summaries of articles for internal use, produce user manuals based on technical manuals for internal use, produce quizzes ... ChatGPT
- Transcription of interviews - Zoom recordings to text transcriptions
- detect AI - use tools to detect AI use, e.g. in applications, QuillBot
- Generating slides - text to slides tools
- Prototyping
- Social media
- Translation - DeepL, ChatGPT
- Image post-production (light, shadows) - Canvas
- Generate captions -
- Coding
- Data analysis - to analyse anonymous data (e.g. visitor numbers, questionnaire data)
Next meeting - decide how to role out statement across the lab
Follow up with Nick Ziogas from KT to link up with everything else happening with AI (Nick organised the workshop in AI: https://indico.cern.ch/event/1352021/
There are minutes attached to this event.
Show them.