Gary Ackerman @garyackerman@qoto.org

AI isn't about mimicking human intelligence; it's about computers as cognitive prostheses, doing things humans can't or wouldn't, like processing vast amounts of data. #AIinEducation

Select response tests inherently have false positives and negatives. They can test game-playing skills rather than true understanding, lacking "construct validity". #TestingProblems #ConstructValidity

A blog post, now with audio:
https://hackscience.education/2018/05/11/non-neutrality-of-technology/

Science and Pseudoscience
https://hackscience.education/2025/05/08/science-and-pseudoscience/

Over time, the ZPD "goes up". Problems once "too complex" move into the ZPD, then into the "too easy" category.

For decisions with significant consequences, ensure a human is in the loop. AI should act as a copilot, assisting humans who make the final critical decisions. #HumanInTheLoop #AICopilots

Generative AI. The concerns are no different than the concerns raised about every other technology.

Deploying GenAI in enterprise comes with risks. Be cautious of hallucination, bias amplification, and misuse. Responsible AI practices are crucial for mitigating these pitfalls. #AIEthics #GenAIRisks

Problem-centered learning breaks down complex knowledge into teachable skills, provides diverse strategies for presenting and practicing these skills within authentic contexts, and emphasizes learner control, clear organization, effective use of multimedia, and consistent feedback

"Providing learners with control over their learning path and making the course structure and content organization transparent. " I wonder what it'd be like to organize school in this way.

Using multimedia can enhance demonstrations, for example, with successive disclosure of text/graphics synchronized with audio or animated graphic devices to focus attention. Keep information and portrayal concurrent. #EducationalTechnology

The Problem-Centered Principle focuses instruction on real-world problems. This can be implemented via a problem progression, where learners solve increasingly complex versions of a problem. #ProblemBasedLearning

Learning is promoted when learners observe a demonstration of the knowledge & skill to be learned. "Tell" presents information, while "Show" demonstrates with visuals. #LearningPrinciples

Demonstration (providing content) & Application (requiring learner response). This framework supports better learning outcomes. #elearning

Some fields labeled pseudoscientific primarily offer negative arguments against established science rather than developing their own positive, testable proposals. Scientific status depends partly on positive contributions.

Pseudoscientists have ready explanations for any observational result, turning apparent counterevidence into support. Science tolerates failures of fit but uses them to refine theories

Pseudoscience often finds confirming evidence everywhere, making its claims essentially unfalsifiable. In contrast, genuine science seeks to falsify hypotheses.

Thomas Kuhn challenged traditional views by emphasizing the history of science. He introduced the concept of paradigms and argued that rival paradigms can be incommensurable, lacking a neutral standard for comparison.

Karl Popper proposed falsification as the criterion to distinguish science from pseudoscience. Unlike positivists, he denied that scientists seek to confirm theories; genuine science seeks to refute hypotheses.

I’ve come to enjoy the reaction of adults (including teachers, principals, etc., parents, business people...) who learn the purpose of public education is not to create people just like them. #edchat #education #school