A survey published this month, commissioned by IBM and conducted by Morning Consult across more than a thousand educators and a similar number of parents, found that a large majority of educators, roughly four in five, feel confident they can teach students about AI: what it is, how it works, and its limitations. A similar majority of parents reported confidence that schools are ready to handle that instructional responsibility well. On the surface, this reads as genuinely encouraging news for a school year that's opened with real anxiety about AI in classrooms.

It's worth being precise about exactly what this confidence covers, though, since "confident teaching about AI" and "confident using AI as an instructional tool," including for grading, are genuinely different competencies. A teacher can feel well equipped to lead a lesson explaining how a language model works, its tendency to produce confident-sounding but sometimes inaccurate answers, and appropriate versus inappropriate student use, while still feeling genuinely uncertain about whether or how to use an AI-assisted grading tool in their own weekly workflow.

This distinction matters because AI literacy instruction and AI-assisted classroom tools, grading support among them, have been developing along somewhat separate tracks. A district can invest heavily in AI literacy curriculum for students, which several recent district announcements this fall have emphasized, without necessarily having equipped teachers with training or clear guidance on using AI tools for their own instructional or assessment tasks.

Where teacher confidence still genuinely lags

Other recent survey data from education research organizations tracking AI adoption in K-12 settings paints a more mixed picture specifically around teachers' own tool use, as opposed to teaching about AI conceptually: professional development specifically covering how to use AI tools in one's own practice, including grading, has grown substantially over the past two years but still hasn't reached most teachers nationally, and many who have received some training describe it as introductory rather than deep enough to confidently evaluate a specific tool for their own classroom.

• Distinguish between confidence teaching about AI conceptually and confidence using AI tools in your own grading workflow
• Check whether your school's AI professional development has covered instructional tool use specifically, not just AI literacy content
• Ask colleagues who've used AI-assisted grading tools directly about their actual workflow, rather than relying on general AI literacy training
• Look for vendor-provided onboarding and training when adopting any new AI-assisted grading tool, since general AI literacy doesn't automatically transfer
• Recognize that strong survey confidence about teaching AI concepts doesn't necessarily predict comfort with specific classroom tools

What this means for schools planning professional development this year

Schools investing in AI professional development this year have a real opportunity to address both tracks deliberately rather than assuming strong AI literacy confidence covers instructional tool competence too. Training that helps teachers explain AI concepts to students is genuinely valuable and, based on this month's survey, appears to be landing well. Separate, more hands-on training specifically covering how to evaluate and use AI-assisted instructional tools, including grading support, in a teacher's own actual weekly workflow remains a distinct and, based on other recent survey data, still underdeveloped need.

Departments considering an AI-assisted grading tool this year should factor this distinction into their rollout planning: strong general educator confidence about AI doesn't substitute for tool-specific training on how a particular grading tool actually integrates with an existing rubric and workflow.

The bigger picture this survey points to

Taken together with other recent data on the state of AI professional development, this month's survey suggests genuine progress on one front, general AI literacy confidence among educators, alongside a real, ongoing gap on another: hands-on competence with specific instructional AI tools. Recognizing both trends separately, rather than treating rising AI literacy confidence as evidence the whole adoption challenge is resolved, gives schools a more accurate picture of where their own professional development investment is still most needed.