What You Should Know
CSU Libraries have developed a comprehensive AI Literacy guide that serves users at every level, from beginners to advanced users. The guide covers essential topics including AI fundamentals, teaching and research applications, prompt engineering, and more. The resource is consistently updated to reflect relevant news, information, and strategies to ensure you are always equipped with the most up-to-date information.
AI In Education
- AI for All Courses: University of Mary Washington joins the growing number of universities that are offering official courses on introduction to AI. The course is asynchronous, pass-fail, and free to current students. These AI courses focus not only on technical understanding, but also ethics and real-world AI applications.
- AI to Support Teaching and Learning: University of Michigan has funded 10 projects to explore GenAI’s potential in teaching and learning. Projects include developing tools to mitigate bias in medical training performance evaluation, using AI to provide targeted feedback on assignments, and personalized dental tutoring.
- Faculty Struggling to Keep Pace: Educators and leaders shared insights at University of Baltimore’s recent AI Summit. Key takeaways include:
- While AI skills are becoming essential across all fields, “human” skills like versatility, leadership, teamwork, and communication are becoming even more sought after.
- Faculty are overwhelmed with how fast the AI space is moving, citing how to adapt assessment as one of the most difficult issues to tackle. A “Process over Product” approach was frequently expressed, although what a successful model would look is still elusive.
AI In Research
- Hidden Prompts for Positive Reviews: Researchers from 14 academic institutions were found to have included hidden prompts across 17 preprints on arXiv. The prompts were hidden using methods like white text or small font sizes, and designed to fool AI-based review systems into giving only positive reviews, with one researcher defending it as “a counter against ‘lazy reviewers’ who use AI”.
- Training AI To “Behave” Like Humans: Researchers introduced Centaur, a foundation model designed to predict and simulate human behavior. The model was fine-tuned on a dataset of over 60,000 participants’ performance across 10,000,000 choices in 160 experiments. It outperformed both the base model and domain-specific cognitive models, marking a significant advancement in computational cognitive science that can be used for “model-guided scientific discovery”.
AI Current Events
- Senate Removes AI Regulation Provision: The U.S. Senate removed an AI provision in the One, Big, Beautiful Bill that would have prevented states from regulating AI for 10 years. Supporters of the provision argued that it would prevent a patchwork of state AI laws hindering AI advancement, while opponents asserted that it would just allow AI companies to avoid accountability in pursuit of innovation.
- NYT Begins Searching Deleted Chats: After winning a lawsuit against OpenAI, the New York Times will begin searching ChatGPT chat logs for evidence of copyright material that they assert the AI company illegally used to train its models. OpenAI contends that the decision undermines the data privacy of its userbase.
Weekly AI Tip
If you are giving AI a particularly complex task or large amount of context, end your prompt with “Before you begin, do you have any questions?”. This often helps to catch any gaps in the context you have provided or misinterpretations that may steer the model down the wrong path.
AI Spotlight
Dr. Robert Ghrist, professor of mathematics and engineering and associate dean of undergraduate education at the University of Pennsylvania, is an educator that is actively and intentionally integrating AI into his courses. In February he gave a great example of how he has incorporated AI tools into an applied linear algebra course where he:
- Created his own course text utilizing Claude, where the AI platform wrote most of the content but it was directed and reviewed by Dr. Ghrist.
- Used the text to power a custom GPT which acted as an AI tutor, which students had access to 24/7. The AI tutor tracks where in the course the student is and will align its responses with the appropriate learning objectives.
- Records the audio of all lectures, which can then be uploaded to NotebookLM to help generate lecture notes, summaries, podcasts, practice exam questions, etc.
- Assigns very difficult weekly homework assignments (co-created with Claude) worth 30% of their grade. The homework acts as a “marshmallow test”; Dr. Ghrist knows that students could just use AI to get all of the points, but students will likely fail the midterms and final. The onus is on the students to use AI in a collaborative way by asking questions/clarifications to ensure they fully understand the material.
Are you using AI in innovative ways and want to share with your colleagues through this newsletter? Email Brian Kelly at [email protected]
Upcoming AI Events
Building Custom GPTs & Gems Workshop – August 8, 10:00AM-2:00PM in DMC 101
- The Building Custom GPTs and Google Gems Workshop focuses on creating personalized and functional AI tools using ChatGPT and Google Gemini. Participants will learn to develop AI tutors, AI generators for assessment content, and AI writing/editing assistants. The workshop is designed so each attendee leaves with their own functional custom AI and the foundational knowledge to build additional tailored solutions. RSVP
