Presenters: Joy Chen, GSE Entrepreneur-in-Residence; Auston Stamm, Digital Accessibility Instructional Specialist; Reuben Thiessen, Project Management Specialist, Stanford Accelerator for Learning
Recording of the session:
- How can AI tools enhance workflows beyond just chatbots?
- What are some key AI tools for accessibility, project management, and education?
- What are some important considerations when implementing AI tools?
“You’ll hear [that AI-generated captions have] around 90% accuracy or so. But…think about that, that actually means that [you’re missing] one out of every 10 words!” – Auston Stamm (12:22)
“[Arxiv.org] had 16,000 paper submissions just last month, [August 2023]! You can see [the number of submissions] is starting to look like an exponential curve upwards. It’s not slowing down. It’s just more and more information coming out. And so what can we do to process and prioritize the information?” – Reuben Thiessen (2:58)
“While AI is incredibly powerful, it still needs humans, such as educators and administrators to establish some guidelines, limits, and boundaries regarding usage.” Joy Chen (35:13)
“Generally AI can be our ‘friend’ and can also be our ‘enemy.’ It can be helpful and also can be harmful. It can definitely enhance learning experiences, but it also raises concerns regarding academic integrity, data, privacy, information, accuracy, and equal access, and also if the student overly depends on AI generated content, they might not develop [essential] critical thinking skills.” Joy Chen (33:55)
- AI summarization tools can help digest large amounts of text and extract key information
- Tools exist which are useful for not only summarizing research papers and newsletters, but also podcasts and videos
- Consider the possibility of accidental plagiarism if publishing AI-generated summaries
- AI-generated captions require manual editing for accuracy, because they often can’t figure out intent. However, they can provide a good starting point for transcriptions editing
- AI-generated alt-text descriptions for images are not always fully accurate because they don’t take into account the author’s intent
- AI tools can help convert documents and add structure to scanned documents
- AI can generate project plans to use as starting point for project types that may be new to you
- Adding metadata (attendees, date, meeting purpose, etc.), as you take meeting notes can help AI later make sense of them
- Consider privacy if you copy-paste your meeting notes to a consumer-facing LLMs such as ChatGPT or Google Bard
- We still need humans in the loop to establish guidelines and oversight in using AI ethically
- We’re living within several AI paradoxes:
- AI can be both helpful and harmful – it has potential to enhance learning but also raises concerns around ethics, accuracy, and access
- AI is powerful yet still relies on human oversight – it can generate content but humans are needed to validate quality and appropriateness
- AI provides accessibility yet can also be restrictive – it breaks geographical barriers but language and digital literacy issues persist