The Future of ABA and AI: An Honest Conversation From Our Team

It’s hard to go a single day without encountering a conversation about artificial intelligence. From the tools we use to write emails to the algorithms shaping what we see online, AI is woven into the fabric of modern life — and the field of Applied Behavior Analysis is no exception. At Into Inclusion NY, we’ve been having this conversation internally for a while now. And we want to share it with you honestly, because we think you deserve more than a polished press release.

This post is written primarily by Kathleen Penland, but it reflects both founders’ perspectives — including the places where Caitlin Bennett and Kathleen genuinely don’t see eye to eye. We think that’s worth naming upfront, because the disagreement itself is meaningful.

First, a Quick Word on ABA

For parents, caregivers, and community members who may be newer to this world: Applied Behavior Analysis is a scientific approach to understanding and changing behavior. It’s grounded in decades of research and is one of the most evidence-based interventions available for autistic individuals and others with developmental differences.

But ABA is not just a set of techniques. At its best, it is a deeply relational practice. The therapeutic relationship — the trust built between a learner and their clinician over time — is not incidental to the work. It is the work. A skilled behavior analyst learns to read a child’s subtle cues, to adjust in real time, to celebrate small victories that might look invisible to an outside observer. That human attunement is irreplaceable. It’s why we do this work, and it’s the lens through which we evaluate everything — including new technology.

Where We See Genuine Promise

Let’s start with the areas where there’s more openness on our team — and where Kathleen, in particular, sees real potential.

Administrative and operational tasks are an obvious starting point. Scheduling, documentation, billing, insurance authorizations — these are time-consuming, often tedious, and pull clinicians away from the work they trained to do. If AI tools can reduce that burden meaningfully and securely, that’s time returned to families. That matters.

Program development and data analysis is another area of genuine interest. ABA is inherently data-driven. We collect data constantly — on skill acquisition, on behavior, on generalization across settings. AI tools that can identify patterns across large datasets, flag when a program isn’t producing the expected outcomes, or surface insights that a busy clinician might miss — these could genuinely strengthen the quality of care, when used thoughtfully and under clinical supervision.

Caregiver support tools represent a third area of promise. Families often need accessible resources between sessions — visual supports, communication aids, plain-language explanations of what their child is working on and why. AI-assisted tools that help bridge that gap, making information more accessible and responsive, could be a meaningful addition to a family-centered model of care.

These are the areas where we’re most open to exploration. They share something in common: AI is supporting the human doing the work, not replacing the human relationship at the center of it.

Where We Have Serious Concerns

Now for the harder conversation — and this is where Caitlin’s voice is especially present.

The therapeutic relationship cannot be automated. We said it above, but it bears repeating. The moment a child looks up at their therapist and tries something new because they feel safe — that is not a data point. It is a human moment. No algorithm can replicate it, and we should be deeply cautious about any technology that positions itself as a substitute for it.

Bias in AI models is a real and documented risk. AI systems learn from the data they’re trained on. If that data doesn’t represent the full diversity of learners — across race, ethnicity, language, disability profile, socioeconomic background — the recommendations those systems produce will reflect those gaps. In a field that already has a complicated history with equity and access, this is not a theoretical concern. It is an urgent one.

Data privacy and HIPAA compliance are non-negotiable. Our clients are among the most vulnerable members of our community. Their behavioral data, their medical histories, their family circumstances — this information is sacred. Any AI tool operating in a clinical context must meet the highest standards of data security and legal compliance. The landscape of AI products is moving faster than regulatory frameworks can keep up, and that gap is a risk we take seriously.

Over-reliance on algorithmic recommendations is a clinical and ethical danger. Every learner we work with is a unique human being. Algorithms are, by definition, generalizations. There is a real risk that clinicians — especially those under pressure, with large caseloads — could begin to defer to AI-generated recommendations in ways that flatten the individuality of the people they serve. That would be a profound failure of our professional responsibility.

Our Ethical Grounding

The BACB Ethics Code — the professional and ethical framework that governs behavior analysts — is clear on the principle of beneficence and non-maleficence: we must act in the best interest of our clients and, above all, do no harm. As we navigate questions about AI, this is our north star. New tools must be evaluated not by their novelty or efficiency alone, but by whether they genuinely serve the wellbeing of the individuals and families we support.

We are not the only ones in the field wrestling with this. Jennings and Cox (2023), writing in Behavior Analysis in Practice, have begun the important work of examining how AI intersects with ABA practice and ethics. Their work is a reminder that this is a field-wide conversation — one that is just beginning, and one that needs many voices.

Two Founders, One Honest Conversation

So where does that leave us? Caitlin is skeptical of AI use in ABA in a meaningful way — she worries that even well-intentioned adoption of these tools risks eroding the relational core of the work and opening doors to harms we can’t fully anticipate. Kathleen is more cautious than enthusiastic when it comes to direct clinical settings, but genuinely excited about what AI might offer on the administrative, operational, and program development side of the house.

We don’t think this disagreement is a problem. We think it’s a feature. The field of ABA needs people asking hard questions and people willing to explore new possibilities — and it needs them in conversation with each other, not in separate camps. That’s what we’re trying to model here, imperfectly and honestly.

This Conversation Belongs to All of Us

We’ve shared our thinking. Now we want to hear yours.

If you’re a parent or caregiver: What would you want to know before AI tools were used in your child’s care? What would make you feel safer — or more concerned?

If you’re a fellow practitioner: How are you thinking about this in your own work? What guardrails feel essential to you?

At Into Inclusion NY, our mission is to build inclusive spaces where all families can thrive. Our vision is a community where belonging isn’t an afterthought. That vision doesn’t change because the technology landscape does. But it does mean we have a responsibility to engage with these questions carefully, transparently, and together.

We’d love to hear from you. Reach out, comment, share this with someone who’s thinking about these questions too. The best conversations about the future of this field won’t happen in boardrooms or tech conferences — they’ll happen in the communities we’re all trying to build.