Supervision Uncovered: Navigating NYS Regulations, Unrestricted Hours, and AI in Fieldwork Training

Overview & Clinical Significance

Supervision is often described as the cornerstone of ethical behavior analyst training, yet the regulatory and practical landscape that supervisors and trainees must navigate has become significantly more complex. In New York State and in many comparable jurisdictions, the intersection of BACB certification requirements, state licensure regulations, scope-of-practice definitions, and emerging technology introduces a level of complexity that many practitioners — experienced and new alike — are not adequately prepared for.

This course examines that complexity directly, focusing on three interconnected areas: the core elements of high-quality supervision in New York State educational contexts, the implications of recent NYS regulatory changes for supervisors and trainees, and the ethical integration of AI tools into foundational behavior analytic training. Each of these areas carries real consequences for trainees accumulating supervised fieldwork hours and for supervisors who sign off on those hours.

The clinical significance of understanding the regulatory environment is not merely bureaucratic. Supervisors who are unaware of how NYS regulations affect which activities count toward unrestricted fieldwork hours may inadvertently sign off on hours that are later disqualified during licensure review. Trainees who accumulate hundreds of supervised hours under a misunderstanding of what those hours require may face significant professional setbacks at the moment of licensure application. Getting this right from the start protects both parties and, ultimately, the quality of the practitioners entering the field.

The AI dimension adds a layer of complexity that is genuinely new. As AI tools become more accessible and are increasingly integrated into clinical workflows, supervisors must think carefully about how to introduce trainees to these tools in ways that develop foundational competencies rather than circumventing them — and must navigate significant uncertainty about how regulatory bodies will treat AI-assisted practice.

Background & Context

New York State's licensure requirements for behavior analysts were established under the Behavior Analyst Licensing Act, which went into effect in 2016 and has been subject to revisions since. The NYS licensing framework requires supervised fieldwork hours that meet specific criteria for supervisor qualifications, supervision format (individual vs. group), and the nature of the activities that may be counted toward restricted versus unrestricted hours. These criteria differ in important ways from BACB certification requirements, creating the potential for compliant-BACB, non-compliant-NYS supervision arrangements.

The unrestricted hours designation in NYS refers to fieldwork activities that span the full range of BACB task list competencies, as opposed to restricted hours which are more narrowly defined in scope. Trainees seeking to maximize the value of their fieldwork must understand how activities are classified and whether their supervisors and practice settings provide opportunities for activities that count toward unrestricted hours. In educational program settings — where many NYS trainees accumulate hours — the range of activities can be narrower than in clinic settings, creating gaps that must be proactively addressed.

Recent regulatory updates in NYS have addressed several areas of prior ambiguity, including supervision delivery methods, documentation requirements, and the qualifications of individuals who may provide supervision for licensure purposes. These updates have practical implications for programs that had been operating under interpretations of earlier language that are no longer supported by current regulation.

The introduction of AI tools into clinical practice — including AI-assisted session note generation, behavior plan drafting tools, and AI-powered data analysis — has created new questions about trainee development. The concern from a training perspective is that premature or unsupervised use of AI tools may prevent trainees from developing the foundational analytic and writing skills that competent behavior analysis requires. Supervisors have an active role in deciding how AI is introduced, under what conditions, and with what scaffolding.

Clinical Implications

For supervisors in NYS educational programs, the clinical implications of current regulatory guidance center on documentation specificity and hour classification accuracy. Supervisors must be able to identify, for each supervision session, which activities were included and how each maps to either restricted or unrestricted hour categories under NYS regulations. This requires supervisors to keep detailed records that go beyond the summary-level documentation that many were accustomed to maintaining.

Trainee role clarity is a related implication. NYS regulations and BACB guidelines both specify conditions under which trainees may work independently and conditions requiring direct supervisor presence or co-implementation. In educational settings where trainees often function as paraprofessionals or behavior technicians alongside their fieldwork activities, the boundary between technician role and trainee role must be explicitly managed. Time spent in a technician role without supervisor oversight does not count toward fieldwork hours, and confusion about this boundary can lead to documentation errors.

The competence domain question — ensuring that trainees are supervised in areas where the supervisor has demonstrated competence — is especially relevant in NYS programs that operate across diverse school settings. Supervisors in educational ABA contexts may be highly competent in skill acquisition and behavior support in academic environments but have limited experience with the medical, behavioral health, or communication-related domains that may arise in certain student profiles. Ethics Code 5.02 requires supervisors to be honest about these boundaries and to establish supplementary supervision or consultation arrangements when a trainee's learning needs exceed the supervisor's competency scope.

For AI tool integration, the clinical implication is the sequencing of introduction. Trainees who encounter AI writing or assessment tools before they have developed independent fluency in clinical writing, behavior plan development, and data interpretation risk developing dependency on AI scaffolding rather than competency. Supervisors should have explicit policies about when and how AI tools may be used in trainee work products, including requiring unassisted demonstrations of competency before AI-assisted versions are permitted.

Ethical Considerations

BACB Ethics Code section 5.01 requires supervisors to provide supervisees with accurate information about the supervision arrangement, including its parameters and any limitations. In the NYS context, this provision has specific operational meaning: supervisors must inform trainees about which activities will count toward which categories of hours, what documentation is required, and what the supervisor's regulatory obligations are within the state's licensing framework. Trainees who make employment and educational decisions based on inaccurate information about how their hours will be classified suffer direct professional harm.

The areas-of-competence obligation under Ethics Code 5.02 applies with particular force in educational program settings where the population and service context can vary significantly from the supervisor's primary training experience. Supervisors who accept trainee placements in settings they are not competent to supervise — even with good intentions — are making an ethics decision, not merely a logistical one. The solution is not to decline all opportunities but to be transparent about competency boundaries and to establish additional supervision or consultation to fill the gaps.

AI tool use raises specific ethics considerations that the BACB has begun to address in guidance documents. The core concern from an Ethics Code perspective is that the use of AI tools must not diminish the quality of client services (Ethics Code 2.01) or misrepresent the authorship or basis of clinical documentation. A behavior plan submitted as the trainee's work but substantially generated by AI without attribution or competency verification raises questions about academic and professional honesty that supervisors must navigate proactively.

Finally, the advocacy function of supervisors and educators is relevant here. NYS regulations have an outsized effect on access to supervision and, therefore, on access to the profession itself. Supervisors who identify regulatory requirements that are creating unintended barriers for trainees — particularly trainees from underrepresented communities or non-traditional educational pathways — have both a professional and an ethical basis for participating in regulatory advocacy processes.

Assessment & Decision-Making

Assessing whether a supervision arrangement meets NYS requirements involves a structured checklist review that goes beyond BACB standards verification. Supervisors should audit their current documentation against current NYS licensing board guidance at the beginning of each supervisory relationship and annually thereafter, as regulations have changed and may continue to change. This audit should address supervisor qualifications, documentation formats, supervision hour ratios, and activity classification procedures.

For trainee hour classification, a decision tree approach helps supervisors consistently apply the restricted-unrestricted distinction across diverse fieldwork activities. Key decision points include: Was the activity directly related to an existing behavior analytic service? Was the supervisor present or available in the manner required? Is this activity explicitly listed or excluded in current NYS guidance? Documenting the reasoning behind each classification in supervision records provides protection in the event of a later dispute about hour validity.

For AI tool integration decisions, supervisors should apply a competency-first framework: the trainee must demonstrate the ability to perform a clinical task independently at criterion before using AI tools to support or augment that task. This decision rule parallels the prompting hierarchy logic familiar from skill acquisition programming — full independence is established before prompts are introduced, not the reverse. Applied to AI, it means the trainee writes a behavior plan before using an AI drafting tool, analyzes a data set before using an AI analysis tool, and writes a session note before using an AI summary tool.

For supervision quality assessment more broadly, structured observation of supervision sessions by external evaluators — a process analogous to inter-observer agreement for clinical observation — provides the most valid evidence of supervision quality. Programs that assess supervision quality only through trainee satisfaction surveys or completion checklists are not capturing the behavioral dimensions that determine whether supervision is actually developing the trainee's clinical competencies.

What This Means for Your Practice

If you are supervising in NYS or in any state with a specific licensing framework that differs from BACB standards, the most important immediate action is a direct review of current licensing board guidance — not a secondhand summary, but the actual regulatory language. Licensing board requirements are updated more frequently than many practitioners realize, and operating on the basis of guidance received in graduate school or during initial licensure may mean working within an outdated regulatory understanding.

For supervisors who have accepted trainees in educational program settings, review your current documentation practices against current NYS guidance for activity classification. If there is ambiguity in how you have been classifying unrestricted versus restricted hours, contact the licensing board directly for clarification and document their response. Proactive clarification is far less costly than discovering a classification error after hundreds of supervised hours have accumulated.

For programs and clinical directors developing policies on AI tool use in trainee work, create a written policy that specifies competency prerequisites for AI tool use, how AI-assisted work products should be labeled and reviewed, and what supervisory oversight is required when AI tools are part of the workflow. A written policy protects both the trainee and the supervisor and signals to trainees that AI use in professional contexts requires the same deliberate, competency-grounded approach that all clinical tools require.