AI Tools So Powerful Universities Ban Them

Key Concepts:

• AI Writing Tools: Tools that assist in generating text, augmenting thoughts, and improving writing.
• Auto Writers: Tools that suggest and generate text as the user writes, like Jenny AI and Yomu.
• Done-For-You Tools: AI tools that can generate entire literature reviews, paper drafts, and systematic reviews with minimal user input, such as Thesis AI and Gatsby.
• Agentic AI: AI systems that can perform complex tasks autonomously, like Manis and Genspark.
• Systematic Reviews: Comprehensive reviews of existing literature on a specific topic, often automated by tools like Elicit.

1. AI Writing vs. Auto Writers

• AI Writing: Involves using tools like ChatGPT, Claude, Perplexity, and Gemini to help augment thoughts and improve writing. This is generally accepted.
• Auto Writers: Tools like Jenny AI and Yomu generate text suggestions as the user writes. The user retains control over accepting or rejecting these suggestions.
  • Example: Jenny AI provides grayed-out text that the user can accept, including references.
  • The speaker notes that the "gray zone" arises when users relinquish control and allow the AI to write without oversight.

2. Done-For-You Tools: Thesis AI

• Functionality: Thesis AI generates in-depth, fully referenced literature reviews on various topics.
• Process:
  1. The user specifies the number of papers to include (e.g., 30 ± 5).
  2. The user can upload references or use Semantic Scholar to find them.
  3. Thesis AI generates a detailed literature review (e.g., 44 pages) in about 20 minutes.
• Output: The output is fully referenced, detailed, and customizable, and can be exported to Word or LaTeX (Overleaf).
• Significance: The speaker highlights that universities and journals are wary of such tools due to their potential for misuse.

3. Done-For-You Tools: Gatsby

• Functionality: Gatsby helps discover research ideas, write scientific papers, patent inventions, and run meta-analyses.
• Process: The speaker created a first paper draft from ideas in a Word document using Gatsby.
• Significance: Gatsby simplifies the creation of initial paper drafts, allowing researchers to identify gaps and refine their narratives.

4. Agentic AI: Manis and Genspark

• Functionality: These tools can generate paper drafts from figures and data.
• Manis:
  • The speaker provided figures to Manis and requested a paper draft.
  • Manis generated a formatted academic paper with sections like experimental methods and results/discussion.
  • While not fully referenced, it provides a strong starting point for peer-reviewed writing.
• Genspark:
  • The speaker provided figures and asked for a story structure for a peer-reviewed paper.
  • Genspark generated a paper draft, including figures in the correct sections.
  • It identified acronyms (e.g., silver nanowires, carbon nanotubes) from the figures.
• Significance: These tools can automate tasks like generating first drafts and providing feedback, but universities and journals are hesitant due to concerns about misuse.

5. Elicit and the Shifting Landscape

• Functionality: Elicit allows users to write reports, search papers, conduct deep reviews, create literature reviews, and even generate PowerPoint presentations.
• Systematic Reviews: Elicit can perform complete systematic reviews with a few clicks and prompts.
• Concerns: While researchers are excited about the time-saving potential, universities and journal editors are wary due to the unknown implications.
• Argument: The speaker argues that tools like Elicit, currently in a "naughty red box" of restricted tools, will eventually be accepted as academia embraces the full power of AI.

6. Notable Quotes:

• "These are the tools that journals and universities quietly draw the line at."
• "...universities uh journal editors are scared because they just don't know."
• "That is going to be the future of academia and all of these tools that are in the red box of do not use will slowly slowly be allowed..."

7. Conclusion:

The video explores a range of AI tools that can automate various aspects of academic research and writing, from generating literature reviews to drafting entire papers. While these tools offer significant time-saving potential, they are met with resistance from universities and journals due to concerns about misuse and the unknown implications of AI in academia. The speaker argues that these tools will eventually be accepted as academia embraces the full power of AI, transforming the future of research.