Frequently Asked Questions: The Ethics of Not Using AI in ABA Practice

1. How can avoiding AI be an ethics violation?

BACB Ethics Code (2022) Code 2.01 requires behavior analysts to rely on current scientific knowledge and to provide effective services. Code 1.01 requires staying current with developments in the field. If AI tools have demonstrated value in specific ABA applications—improving documentation accuracy, assessment synthesis, or supervision delivery—then declining to evaluate those tools without examining the evidence may produce worse client outcomes than thoughtful adoption would.

Kaur et al. (2026) showed that restricting assessment to a narrow tool set misses clinically relevant information. The same logic applies to practice technology.

2. Does this mean all BCBAs are required to use AI?

No. The argument is about symmetric evaluation, not mandatory adoption. The same evidence-based reasoning practitioners apply to intervention selection should be applied to technology adoption decisions.

Rejecting AI without examining the evidence is no more defensible than adopting it without examining the evidence. Practitioners who have conducted a structured evaluation of AI tools for their specific practice context and concluded that no current tools add value in that context are in a different position from practitioners who have categorically declined to engage with AI at all.

3. What specific AI applications have demonstrated value in ABA?

Applications with the most documented utility include AI-assisted note drafting (reducing administrative time while requiring clinician review and signature), AI-assisted pattern analysis in large behavioral datasets, and AI-assisted literature search for evidence-based practice review. Treviño & Gerstein (2026) validated an emotion dysregulation assessment—AI tools that assist in administering, scoring, or synthesizing validated instruments represent an extension of that kind of measurement precision. Each application requires independent evaluation against the practitioner's specific context and client population.

4. How does Code 2.01 apply to technology adoption decisions?

Code 2.01 requires behavior analysts to use current scientific knowledge as the basis for their services. If the current scientific knowledge includes evidence that AI tools improve specific clinical outcomes in ABA, then practitioners who have not evaluated those tools may be providing services that do not reflect the best currently available practices. The code's requirement is not static—it tracks the evolving state of the science, which now includes an emerging evidence base for AI applications in behavioral health.

5. What is the difference between skepticism about AI and categorical refusal?

Skepticism is epistemically appropriate and requires structured evaluation: specify what evidence would be needed to justify adoption, seek that evidence, and make a data-based decision. Categorical refusal is an attitude that precedes evidence and is not updated by it. The BACB Ethics Code requires evidence-based practice, which means decisions—including technology adoption decisions—should be based on the best available evidence rather than on categorical preferences.

6. How should I evaluate an AI tool before deciding whether to use it?

Define the clinical problem the tool is intended to address. Specify the quality threshold the tool must meet (accuracy, confidentiality, reliability). Pilot the tool on a small sample with your own independent verification of outputs.

Measure whether outcomes meet the threshold. Make an adoption decision based on the data. Kaye et al.

(2025) demonstrated that combining assessment approaches produces more accurate function hypotheses—the same combined-methods logic applies: evaluate AI tools in the context of the specific problems they are intended to solve rather than in the abstract.

7. Can AI improve supervision delivery?

AI tools can support supervision by reducing the administrative burden on BCBAs—time spent on documentation can be reallocated to direct supervision activities. For practitioners with high caseloads, this reallocation can be clinically significant: more time for supervision means more development for RBTs and better service quality for clients. Dawson et al.

(2026) found that communication training outcomes depend on matching the training approach to the individual's profile and environment. Supervision quality shows the same profile-specificity, and AI tools that free up supervisor time for that individualized work may produce meaningful clinical gains.

8. What should practitioners do if their organization prohibits AI use?

Practitioners should understand the specific scope and rationale of any organizational policy before evaluating its ethics implications. Policies that prohibit use of non-HIPAA-compliant tools for client data are ethically sound and should be followed. Policies that categorically prohibit evaluation of AI tools for any purpose may warrant engagement through organizational channels—presenting evidence about specific tools, specific applications, and specific safeguards.

Code 3.01 requires operating within the scope of applicable organizational requirements while maintaining ethical practice.

9. How does the ethics of inaction argument relate to telehealth adoption?

Telehealth provides the most recent precedent: initial resistance based on concerns about fidelity and clinical quality gave way, partly under pandemic pressures, to an evidence base showing comparable outcomes for many applications. The AI cycle appears structurally similar—legitimate initial concerns, categorical resistance in some quarters, and an emerging evidence base that supports thoughtful adoption for specific applications. Practitioners who engaged with telehealth evidence-based practice earlier were better positioned to serve clients during the pandemic; practitioners who engage with AI evidence-based practice now may be similarly positioned for what comes next.

10. What is the ethical obligation when AI could improve services but resources are limited?

Resource constraints are a real variable in practice technology decisions. The ethical obligation is not to adopt every beneficial technology regardless of cost—it is to make adoption decisions transparently, based on the best available evidence, and to advocate through appropriate channels when resource constraints prevent adoption of tools that would meaningfully improve client outcomes. Al Aqel et al.

(2026) found that awareness and resources shape engagement with services. Documenting that a resource constraint—not categorical opposition—is the reason for not adopting a beneficial tool is a different ethical position from declining to evaluate whether the tool would be beneficial in the first place.