Platform
Artificial intelligence is changing what educational technology can recognize, organize, and support.
It may help convert handwriting into searchable content. It may surface a learning pattern hidden across weeks of student work. It may prepare a grading suggestion, organize feedback, or help an educator find a useful learning resource.
These capabilities can save time and reveal valuable context.
They can also be incomplete, inaccurate, biased, or misunderstood.
That is why READTYCH is developing AI around a clear principle:
AI can assist the educational process. People remain responsible for it.
The READTYCH Responsible AI principles guide how AI-assisted capabilities should be designed, evaluated, introduced, explained, monitored, and improved across the Digital Book and connected education platform.
The purpose is not to place artificial intelligence at the center of education.
It is to place better tools in the hands of the people already at its center.
A system can recognize a pattern without understanding the person behind it.
It may see that a student submitted three assignments late. It may not know that the student was absent, shared a device, experienced a technical problem, or needed more time to understand the material.
It may identify an incorrect answer. It may not recognize the reasoning that led to it.
It may suggest a grade. It does not carry the educator’s knowledge of the student, lesson, classroom, and learning objective.
The more capable AI becomes, the more clearly its boundaries must be defined.
READTYCH is building responsible AI governance around the entire life of an AI-assisted capability, from the first product decision to its use in a real educational environment.
That includes asking:
Responsible AI begins before a model produces its first output.
It begins with deciding whether the model belongs in the experience.
Every AI-assisted capability should begin with a defined educational or operational need.
READTYCH does not intend to add AI merely to make the platform appear more advanced. A capability should help a student, educator, administrator, or institution complete a meaningful task more effectively.
AI-assisted outputs should support human judgment rather than quietly replace it.
Educators and other appropriately authorized users should remain responsible for meaningful academic, instructional, administrative, and institutional decisions.
People should be able to recognize when an important output has been generated or materially influenced by AI.
Institutions should receive clear information about what a capability is designed to do, the information it uses, its known limitations, and the role expected of the human reviewer.
AI should not receive student information simply because that information exists within the platform.
Any use of personal or educational information should have an identified purpose, appropriate authorization, suitable safeguards, and a level of access proportionate to the task.
AI-assisted capabilities should be evaluated for patterns that may disadvantage particular students or groups.
Fairness does not mean assuming every system will behave identically in every situation. It means actively examining differences, investigating unexpected outcomes, and improving the system where concerns appear.
An AI-assisted capability should be evaluated against the educational context in which it will be used.
READTYCH should not present an experimental result as a dependable conclusion or an unverified capability as ready for high-impact use.
A meaningful output should provide enough context for an authorized person to understand and evaluate it.
The goal is not always to explain every technical calculation. It is to help users understand what the output means, what may have influenced it, and what they should consider before acting.
AI capabilities, supporting systems, information flows, and external providers should be considered within the wider READTYCH security program.
Powerful tools require careful access, monitoring, testing, and protection.
Responsibility should not disappear inside an algorithm.
READTYCH, participating institutions, authorized users, and applicable service providers should understand their roles in selecting, configuring, reviewing, and using AI-assisted capabilities.
AI performance can change as content, users, models, educational settings, and product capabilities evolve.
Responsible AI requires ongoing monitoring, feedback, testing, and improvement rather than a single review before launch.
Different AI-assisted activities carry different levels of consequence.
Organizing a set of notes is not the same as recommending an academic intervention. Recognizing handwriting is not the same as determining a final grade. Suggesting a resource is not the same as deciding what a student is capable of achieving.
READTYCH’s governance approach considers the possible impact of a capability before deciding how it should be used.
AI may help with tasks such as organizing content, recognizing handwriting, improving search, categorizing materials, or preparing information for review.
These uses may still require privacy, security, and quality controls, but they generally carry less direct academic consequence.
AI may help identify common errors, prepare feedback, suggest resources, summarize learning activity, or surface possible patterns.
The educator should be able to examine the output, consider the context, and decide whether it is useful.
A final grade, disciplinary action, formal student placement, eligibility determination, or significant educational intervention can materially affect a learner.
READTYCH does not intend for AI to make such decisions independently.
The more meaningful the consequence, the stronger the requirement for authorized human review, contextual understanding, documentation, and institutional control.
AI can help bring information forward.
Responsibility stays with the person making the decision.
READTYCH is exploring AI-assisted capabilities across the Digital Book and education platform.
Availability will depend on product development, institutional configuration, technical validation, and the educational setting.
Students may be able to convert handwritten notes, annotations, calculations, and responses into organized or searchable digital content.
Recognition results may contain errors, especially when handwriting, language, symbols, layout, or subject-specific notation is complex. Users should be able to review and correct recognized content where appropriate.
AI may help prepare a possible score, classification, or review suggestion for certain assignments or assessments.
A suggestion should remain a suggestion.
The educator should be able to examine the original student work, understand the relevant criteria, and accept, change, or reject the proposed result.
AI may help educators identify common errors or prepare a starting point for feedback.
The educator should decide whether the feedback is accurate, useful, appropriate, and suitable for the individual learner.
The platform may surface repeated difficulties, changing performance, incomplete work, or areas where additional attention may be useful.
A pattern should invite investigation. It should not become an unquestioned label attached to a student.
AI may help suggest books, lessons, examples, exercises, or educational applications related to a learner’s current work.
Recommendations should reflect approved educational goals and institutional controls. They should not trap students inside a narrow assumption about their abilities.
AI may help organize information, summarize activity, assist support teams, or reduce repetitive platform tasks.
Administrative convenience should still respect access controls, privacy requirements, and institutional responsibilities.
An educator understands more than an assignment response.
They understand what was taught, how the lesson unfolded, what the student has attempted before, what support has already been provided, and what may help next.
AI does not share that relationship.
READTYCH is therefore developing AI-assisted academic capabilities around educator review.
Where appropriate, educators should be able to:
Human oversight should be meaningful.
A person should not be asked to approve an output they cannot understand, question, or realistically evaluate.
Learning intelligence can reveal important changes across time.
It may show that a student is improving steadily, struggling with one concept, missing particular types of assignments, or responding differently to different forms of instruction.
That context can help an educator act earlier.
But a pattern is never the complete student.
READTYCH does not intend to reduce learners to permanent scores, fixed predictions, or automated labels.
AI-assisted insights should be treated as one source of information among many. They should be considered alongside student work, educator observation, classroom participation, learning conditions, individual circumstances, and direct conversation.
A learner’s past performance may provide context.
It should not become a ceiling placed over their future.
Technical transparency alone is not enough.
A long model description may satisfy documentation requirements while leaving the educator who sees the output with no practical understanding of what to do next.
READTYCH aims to provide transparency at different levels.
Students should receive age-appropriate information when AI materially affects their experience. They should understand that the technology can make mistakes and that a person remains involved in meaningful academic decisions.
Educators should understand what the capability is designed to do, what information it considers, what limitations may affect it, and what review is expected before the output is used.
Administrators should be able to understand which capabilities are enabled, who can access them, what information is involved, and how concerns can be reported.
Institutional reviewers may require more detailed information about architecture, information flows, service providers, security, retention, evaluation, and contractual responsibilities.
Larger institutions may need a system-level view of how AI is governed, monitored, configured, and communicated across participating schools.
Transparency should help each person make the decision appropriate to their responsibility.
AI can only produce an output from the information it receives.
That makes data governance part of AI governance.
Before student or institutional information is used by an AI-assisted capability, READTYCH’s approach considers:
Information collected for one educational function should not silently become training material or input for an unrelated AI purpose.
Where AI-assisted features involve student information, the use should be defined, limited, protected, and communicated appropriately.
An AI system may perform differently across languages, handwriting styles, grade levels, subjects, response formats, educational backgrounds, or student populations.
These differences may not be obvious during an early demonstration.
They become visible through careful evaluation.
READTYCH’s responsible AI direction includes examining:
No evaluation can promise the absence of every possible bias.
Responsible practice means continuing to look for unfair outcomes, taking concerns seriously, and changing the system or its use when the evidence requires it.
An AI-assisted output may sound confident even when it is incomplete or wrong.
That is especially important in education, where one incorrect interpretation can influence how a learner is understood.
READTYCH aims to evaluate AI capabilities according to their intended use rather than relying only on general technical performance.
A handwriting recognition tool should be tested on the kinds of writing students actually produce.
A grading suggestion should be evaluated against relevant educator-reviewed criteria.
A learning pattern should be checked for usefulness, consistency, and the risk of misleading interpretation.
A resource recommendation should be appropriate to the subject, learner, institution, and approved educational environment.
The right question is not simply, “Does the AI work?”
It is, “Does it work well enough for this educational purpose, under these conditions, with these safeguards?”
Some uses of AI may conflict with the role READTYCH wants technology to play in education.
READTYCH’s governance direction is built around boundaries such as:
Additional restrictions may be established as the product develops, institutional requirements become clearer, and new AI capabilities are evaluated.
A responsible AI program should be willing to say where AI does not belong.
Responsible AI is a lifecycle, not a final approval screen.
The proposed capability begins with a clear problem, intended user, and expected educational value.
READTYCH considers whether AI is suitable for the task, what could go wrong, who could be affected, and whether a simpler approach would work.
The team considers what information is required, how it will be accessed, whether a provider is involved, and what privacy and security controls are needed.
The product experience should make it possible for the appropriate person to review, understand, correct, reject, or report the output.
The capability should be tested against its intended educational context, including accuracy, reliability, usability, fairness, and possible misuse.
Institutions should receive relevant information and configuration options before a capability is introduced into their environment.
Feedback, error reports, performance signals, and institutional concerns can help reveal issues that were not visible during development.
If a capability does not meet the required standard, READTYCH should be prepared to improve it, narrow its use, temporarily suspend it, or remove it.
Innovation earns trust when it remains accountable after launch.
Some READTYCH capabilities may depend on external technology, infrastructure, models, or service providers.
Working with another provider does not remove READTYCH’s responsibility to understand how that relationship affects schools and students.
Relevant reviews may consider:
Institutions should receive appropriate information about material providers involved in the capabilities they use.
Responsible AI depends on people being able to raise concerns.
An educator may notice an inaccurate grading suggestion. A student may believe an output misunderstood their work. An administrator may identify an unexpected pattern. A privacy or technology team may have questions about how information was processed.
READTYCH intends to support reporting and review processes for concerns involving:
Reported concerns should be evaluated according to their nature, urgency, possible impact, and the information available.
Where appropriate, READTYCH may investigate, correct information, update guidance, adjust a capability, restrict its use, or take other responsive action.
Pilot programs provide an opportunity to evaluate AI-assisted capabilities in real educational settings before wider adoption.
A pilot involving AI should begin with a shared understanding of:
Feedback should include more than whether the feature appeared impressive.
A meaningful pilot should explore whether educators found the capability useful, whether students understood the experience, where errors occurred, whether outputs were fair, and whether the technology improved the learning process without creating an unacceptable burden or risk.
Responsible AI means developing and using AI-assisted capabilities around educational purpose, human oversight, privacy, transparency, fairness, reliability, security, and accountability. AI may support the work, but appropriately authorized people remain responsible for meaningful decisions.
READTYCH is exploring AI-assisted grading suggestions for appropriate educational activities. Any such output should remain subject to educator review. The educator should be able to examine the student’s work and accept, change, or reject the suggestion.
READTYCH does not intend for AI to independently make final academic, disciplinary, placement, or eligibility decisions. AI-assisted learning insights may help authorized educators identify patterns or questions worth exploring, but human judgment and educational context remain essential.
Information used by an AI-assisted capability should have a defined purpose, appropriate authorization, limited access, suitable security, and clear retention expectations. The exact controls will depend on the capability, institutional configuration, and product arrangements involved.
READTYCH’s governance direction includes evaluating whether capabilities perform differently across relevant users, subjects, languages, response formats, and educational settings. Concerns should be investigated, and capabilities may be improved, restricted, or removed when performance does not meet the required standard.
READTYCH intends to provide institutions with appropriate information and controls over the AI-assisted capabilities available within their environment. The exact configuration options will depend on product development, the institutional program, and the capability involved.
The next page starts here
The future of education will include artificial intelligence.
The more important question is what kind of future we choose to build with it.
One where an algorithm quietly replaces educational judgment?
Or one where technology helps educators see more clearly, respond earlier, and spend more time on the human work students need?
READTYCH is choosing the second.
AI can recognize.
AI can organize.
AI can suggest.
AI can help bring important context into view.
The educator understands the learner.
The institution carries the responsibility.
The human makes the decision.
That is where AI belongs in the READTYCH classroom.