What is being claimed, who might benefit, and what remains unresolved? A development earns attention through its educational consequences; novelty alone does not settle its value. Entries appear by event or publication date, newest first; the source-check date is separate.
23 entries
An Intentional Teaching panel considers what computer-science students should still learn when AI can perform parts of their work.
For an instructor: Use the comparison to ask which foundations psychology students need in order to judge a generated explanation.
What this does not establish: Published October 20; recorded during summer. These are practitioner perspectives from another discipline, not a test of a psychology curriculum.
Anthropic: Managed Agents web fetch follows organizational domain restrictions as well as agent-level settings.
For an instructor: Maintain an approved course brief separately from provisional suggestions.
What this does not establish: Domain restrictions constrain access, not the truth of an allowed page. Date basis: Release-note entry.
DeepSeek: DeepSeek’s October17 change log reports the model update and search interface.
For an instructor: Compare the whole application and its permissions, rather than a model name alone.
What this does not establish: A provider’s coding or agent benchmark does not establish better psychology teaching. Date basis: Announced in change log.
OpenAI: Hosted sandboxes add skills for spreadsheets, presentations, and Word documents.
For an instructor: Proposed pilot: delegate a bounded lesson artifact, then inspect conceptual alignment and revision time.
What this does not establish: A finished file still needs content and accessibility review. Date basis: Release-note entry.
Meta: Meta announced weights and several deployment routes under the Muse Model License.
For an instructor: Specify stopping points and approval boundaries for any persistent course workflow.
What this does not establish: Deployment requires suitable hardware, software and license review. No universal laptop or privacy claim follows. Date basis: Release announcement.
A US graduate-student survey reports the greatest AI use in teaching among the domains examined.
For an instructor: Include teaching assistants and graduate instructors in course-level AI planning.
What this does not establish: Use patterns are not effectiveness measures; abstract-only inspection.
Anthropic: In beta, an agent can write a phased program coordinating many agents and combining results.
For an instructor: Maintain an approved course brief separately from provisional suggestions.
What this does not establish: A generated plan may reproduce the same mistaken premise across tasks. Date basis: Release-note entry.
Google: The keynote article describes persistent work, sub-agent coordination, and enterprise governance.
For an instructor: Make the approved source hierarchy and current rubric version explicit.
What this does not establish: Announcement scope is broader than verified general availability. The opened article date is used. Date basis: Visible article publication date.
OpenAI: Agents API adds a browser operated through a managed session.
For an instructor: Proposed pilot: delegate a bounded lesson artifact, then inspect conceptual alignment and revision time.
What this does not establish: Website access approval does not guarantee approval before each consequential action; see computer-use documentation. Date basis: Release-note entry.
OpenAI: Managed sessions support continuing work, tools, and execution environments.
For an instructor: Proposed pilot: delegate a bounded lesson artifact, then inspect conceptual alignment and revision time.
What this does not establish: Public beta is a product status, not an instructional validation. Date basis: Release-note entry.
Google: Google describes generating documents and presentations from selected context across Workspace apps.
For an instructor: Make the approved source hierarchy and current rubric version explicit.
What this does not establish: Eligibility, admin controls, and surface-specific rollout apply. Date basis: Article publication and rollout announcement.
Meta: Meta describes an agent acting through its own virtual computer.
For an instructor: Specify stopping points and approval boundaries for any persistent course workflow.
What this does not establish: The description is a company claim, not a validation of classroom use or a guarantee that every action is safe. Date basis: Launch announcement; page updated September30.
Google: Gemini API video analysis can choose relevant moments and modalities to inspect.
For an instructor: Make the approved source hierarchy and current rubric version explicit.
What this does not establish: This source separates API availability from later app rollout. Benchmark claims do not measure teaching outcomes. Date basis: Article publication.
Two research-methods sections were compared; the trained section did not show a writing-grade advantage.
For an instructor: Evaluate an instructional routine against standard teaching before extending it.
What this does not establish: No blanket conclusion about AI training; restricted full text.
A two-year school experiment reports limited engagement with Khanmigo despite broad access.
For an instructor: Test whether students ask for and use help, alongside measuring achievement.
What this does not establish: School mathematics; combined platform intervention; month-only date.
Two literacy trials separate platform use from reading achievement. Human support increased use, but achievement did not improve.
For an instructor: Track participation and learning as different outcomes.
What this does not establish: Institutional summary; not a psychology-course result; month-only date.
Microsoft: The product-team post describes guided dialogue and practice activities.
For an instructor: Separate an instructor’s preparation workflow from a student’s practice workflow.
What this does not establish: Pedagogical design intentions are not outcome evidence from psychology students. Date basis: Dated vendor announcement.
Anthropic: Anthropic describes separating agent control, persistent history, and execution.
For an instructor: Maintain an approved course brief separately from provisional suggestions.
What this does not establish: Engineering account, not an independent classroom evaluation. Date basis: Article publication.
A 35-study meta-analysis reports positive average outcomes across earlier ChatGPT studies.
For an instructor: Ask whether its tasks and outcomes match the learning you want to assess.
What this does not establish: Underlying studies date to 2022–2024; no Psy101-specific pooled claim.
A TEQSA-hosted Queensland case study discusses criteria, access and coordinated AI expectations.
For an instructor: Review whether your rubric rewards the reasoning your task is meant to reveal.
What this does not establish: Last-updated date; professional advice rather than universal policy.
A Sydney study finds no correlation between reported AI use and grades.
For an instructor: Separate tool use, task performance and evidence of understanding when interpreting results.
What this does not establish: Observational, self-reported use; data concern early 2024.
A review connects assessment validity with integrity processes, marking and institutional support.
For an instructor: Plan staff development and assessment redesign together.
What this does not establish: A review and practice perspective, not an intervention trial.
Microsoft: Microsoft presented Teach as a preparation assistant alongside other education offerings.
For an instructor: Separate an instructor’s preparation workflow from a student’s practice workflow.
What this does not establish: A preparation feature does not by itself establish alignment or reduce the total review workload. Date basis: Announcement.
No developments match. Try another phrase or clear the filters.