Georgia Mandated AI Dyslexia Screening. The Privacy Fight Misses the Real Problem.
Last updated:
When Georgia enrolled your kindergartner in a state-funded AI reading program this fall, you probably did not get a clear explanation of what a flag on the screener means, what the school is now obligated to do about it, or what you should ask for next. Atlanta parents are raising legitimate concerns about whether Amira Learning’s AI tool should be capturing five-year-olds’ voices, phonological patterns, and images. That fight is worth having. It is not the most important fight worth having. The science behind early dyslexia screening is solid. What Georgia built around the screening deserves more scrutiny than any privacy policy.
TL;DR
Georgia's HB 307 named Amira Learning as the state's sole state-funded universal reading screener for K-3 students in 2025-2026, measuring phonological skills in kindergartners before children read words on their own.
Atlanta parents have raised data privacy concerns about kindergartners' voices, images, and phonological data being captured by an AI tool.
APS confirmed Amira complies with FERPA, COPPA, and the Georgia Student Data Privacy Act; opting out is difficult because the state funds the program.
The phonological skills Amira measures are validated by decades of reading science as early predictors of reading risk; the IDA 2025 definition backs early literacy support as "particularly effective."
A screener result is not a diagnosis; parents should ask their district what specific follow-up is tied to the result, and for formal accommodations, a professional evaluation is the required route.
Georgia made AI dyslexia screening mandatory for every kindergartner this fall. Here are the questions parents are asking most, and the one the schools forgot to answer.
Common questions
Is Amira’s AI dyslexia screener accurate?
The skills Amira measures, phoneme segmentation, rapid automatized naming, and nonword repetition, are validated by decades of reading research as reliable early predictors of phonological processing difficulty. No screener catches every child who will struggle, and no screener flags only children who need intensive support. A result is a signal about specific skills, not a verdict about your child’s reading future.
How do I opt my child out of the Georgia dyslexia screening?
Opting out of the AI version is difficult because the state funds the program, but Atlanta Public Schools has indicated it is working to offer a paper-and-pencil alternative for families who request one. Contact your district directly and ask whether that option is available in your school, and get the request in writing.
My child was flagged by the screener. Does that mean they have dyslexia?
No. A screener is a starting point, not a diagnosis. A flag means the tool detected a phonological skill below the risk threshold it was calibrated to catch, which is useful information about where to focus attention. For a formal evaluation, or to pursue an IEP, 504 plan, or other formal accommodations, a professional evaluation is the required route. If you suspect a vision, hearing, or medical cause for reading difficulty, that route also leads to a professional assessment, not a screener result. The screener is the first word, not the last one.
What should I ask the school after the screening?
Ask for the specific scores, not a pass-or-flag summary. Ask what the school’s protocol is for children who are flagged, and what instructional support is available. Ask whether your child will be monitored over time or only screened once. And if you want a fuller picture of how your child learns across the systems reading depends on, ask whether a more comprehensive assessment is available at the school level. The parent who knows what questions to ask is already ahead.
Under HB 307, the Georgia Early Literacy and Dyslexia Act, the state selected Amira Learning as its sole universal reading screener for K-3 students, state-funded at no cost to districts, for the 2025-2026 school year. Amira is an AI program that schools use for both reading instruction and assessment. In kindergarten, where most children are still pre-readers, the screener measures phoneme segmentation, rapid automatized naming, nonword repetition, and visual attention, skills that decades of reading research identify as reliable predictors of future reading difficulty, often before a child reads a single word on their own.
Parents in Atlanta Public Schools have raised concerns since back-to-school season. APS board chair Jessica Johnson acknowledged the questions directly: “Parents have definitely raised some concerns about whether the AI tool that was used as part of the dyslexia screener was leaving, I guess, what they call personal imprints.” Carolyn Boyd, a former teacher with a child entering kindergarten in APS, put it plainly: “We’re giving them our kids’ data, their images, their voices. I don’t feel safe having my kids’ images and voices just out in the ether being used to train AI models to replace the staff that the school should be hiring.”
APS confirmed that Amira complies with FERPA, COPPA, and the Georgia Student Data Privacy Act. Because the state funds the program, opting out is not straightforward. Johnson said: “Because the state is paying for this Amira tool, there are some processes we have to follow in terms of making sure our students are screened.” The district is working with Amira and the Georgia Department of Education to finalize a paper-and-pencil alternative for families who request one.
Author Quote"
We’re giving them our kids’ data, their images, their voices. I don’t feel safe having my kids’ images and voices just out in the ether being used to train AI models to replace the staff that the school should be hiring.
"
What the coverage gets wrong
Most reporting on Georgia's AI dyslexia screener frames this as a privacy-versus-technology debate: parents worried about data, schools defending the tool. That framing misses the structural question that determines whether the screening actually helps any child. The phonological skills Amira measures, phoneme segmentation, rapid automatized naming, and nonword repetition, are the markers the field's own reading science has validated for early-risk detection. Coverage that stops at privacy has not asked what Georgia requires schools to do after a child is flagged, which is the question that separates a screening mandate from an early intervention program.
The screener science is right. The frame around it is the problem.
Here is where coverage goes sideways. The skills Amira measures, phoneme segmentation, rapid automatized naming, and nonword repetition, are the same markers reading scientists have validated for decades as early predictors of phonological processing difficulty. The International Dyslexia Association’s 2025 definition of dyslexia explicitly calls language and literacy support before and during early childhood “particularly effective,” which is the academic case for catching reading risk in kindergarten, before a child has failed two or three grade levels waiting for a threshold that triggers services. The wait-to-fail model that Georgia’s law was designed to replace is a documented harm. Screening young children is, on the evidence, the right call.
The problem is what Georgia built around the screening. A mandate that deploys an AI screener without a guaranteed parent education layer is not an early intervention program. A flag on an Amira report means a child’s phonological processing is below the risk threshold the screener was calibrated for. It does not tell a parent which systems beyond phonological processing the child needs help with, what the school is required to provide in response, or how to interpret the result without turning it into a fixed label. The screener is a door. Georgia opened it without a map of what is on the other side.
The data privacy concern and the post-screening support gap are not the same fight, and they should not compete for the same attention. Reading difficulty draws on several processing systems, and a screener that measures one or two of them in twenty minutes is telling you where to look, not what is there.
Key Takeaways:
1
The screener science is sound: Amira measures phoneme segmentation, rapid automatized naming, and nonword repetition, markers reading research has validated for decades as early predictors of phonological processing risk in pre-readers.
2
The privacy fight is incomplete: Parental concerns about voice and image data are legitimate, but the structural gap, no guaranteed parent education layer after a child is flagged, is the larger problem the mandate did not solve.
3
A result is a starting point: Whether your child is flagged or not, ask the school specifically what the next step is; a screener is not a diagnosis and is not a reading plan.
What to ask before and after the screening
If your child is in Georgia schools, the Amira screening has happened or is happening. Three things are worth doing regardless of the result. First, request the actual output: ask the school for specific scores, not a summary of pass or flag. Second, ask what the school does next. A district that mandated screening and built no response protocol for flagged children has solved the identification problem while leaving the intervention problem untouched. Third, if data privacy is your concern, APS is working toward a paper-and-pencil alternative; ask the school directly whether that option is available in your district.
The screener result is not a diagnosis and is not a reading plan. A flag signals that one or more of the phonological building blocks the screener measures needs attention. That is buildable. A clear result does not guarantee your child is free of reading difficulty; it means the screener did not catch a signal in the systems it measured this round, and there are systems reading depends on that no twenty-minute screen covers. If formal accommodations are on the table, an IEP or 504 plan requires a professional evaluation; the screener is a starting point, not the route to those supports.
The parent who gets the most from this situation is not the one who fights hardest over data privacy alone, or the one who takes the result at face value. It is the parent who asks the school: what do you do with this result, and how do I help my child from here?
Author Quote"
Because the state is paying for this Amira tool, there are some processes we have to follow in terms of making sure our students are screened.
"
Your child’s reading future is not settled by any screener result, and it is not decided by a twenty-minute AI session in kindergarten. The villain here is not the data concern or the technology. It is the pattern of building mandates around identification while leaving the instructional response to chance. You have the right to ask your district for the specific result, the specific follow-up plan, and the specific skills to build at home. If you want to see where your child’s processing actually stands across every system reading depends on, the Learning Success All Access program gives you a full multi-system picture and a personalized plan built for your child, not the average kindergartner.
Is Your Child Struggling in School?
Get Your FREE Personalized Learning Roadmap
Comprehensive assessment + instant access to research-backed strategies
