|
DAILY AI BRIEFING FOR EDUCATORS
HigherEd AI
Monday, December 22, 2025
|
|
|
|
|
Today’s Focus
ChatGPT Gets Fine-Grained Tone Controls
OpenAI rolled out new personalization controls for ChatGPT this week. Beyond the existing style and tone settings, users can now adjust three specific dimensions of how ChatGPT communicates: enthusiasm, warmth, and emoji use. Each setting offers three levels: More, Less, or Default.
This matters more for educators than it might initially appear. Different teaching contexts need different communication styles. When you’re designing a challenging problem set, you might want ChatGPT responses with less enthusiasm and more formality. When you’re creating student-facing explanations, more warmth and moderate enthusiasm changes the tone entirely. With these granular controls, you’re no longer locked into a single voice.
Consider creating tone profiles for different uses: a research-focused profile with less emoji and formal warmth, a student engagement profile with higher enthusiasm and warmth, and a default for general purposes. This transforms ChatGPT from a single tool into multiple tools shaped by your pedagogical intent.
Worth considering:
If an AI’s tone shapes how students perceive and engage with information, should tone control be part of your course design? How might different tones affect student learning outcomes?
|
|
|
Platform News
Anthropic Releases Bloom for Behavioral Evaluation
Anthropic released Bloom this week, an open-source tool designed to evaluate how AI models behave in specific contexts. It allows organizations to test models for safety, accuracy, bias, and other behavioral characteristics before deployment.
For universities building or implementing AI tools, this is important infrastructure. Before deploying any AI system in your institution, you should be running behavioral evaluations. Bloom provides a framework for that testing.
|
|
|
Design Update
Figma Crop Tool Gets Speed Boost
Figma refined its Crop tool this week with faster access via double-click and automatic aspect ratio preservation. While this seems incremental, it reflects how design tools are optimizing for AI-assisted workflows. Faster iteration cycles mean faster experimentation.
For students learning design, this matters. They’re learning tools that are themselves evolving in response to AI workflows. Understanding how to work with tools that anticipate your intent becomes a core skill.
|
|
|
A Final Reflection
When students learn tools that change faster than course syllabi, how do you teach tool mastery versus teaching principles of tool use?
|
|
|
This newsletter synthesizes developments from TLDR AI, TLDR Design, and primary source documentation. Each edition is curated specifically for higher education professionals.
Visit AskThePhD.com for more resources, daily tool tests, and tutorials for educators.
Dr. Ali Green
Professor & AI in Education Specialist
From the AskThePhD team at HigherEdAI
|