AI in Social Work: Staying Fluent While Keeping Your Ethical Guard Up

Most US social workers now encounter AI in their practice. Your MSW should help you stay fluent with the technology and alert to the ethics.

Nearly two-thirds of US social workers employ AI in their current roles. That’s according to a June 2026 survey of 1179 practitioners by the Moritz Center for Societal Impact at the University of Texas at Austin and the National Association of Social Workers (NASW). Many rely on these tools several times a week. That makes literacy a pressing issue.

Use Cases for AI in Day-to-Day Social Work

Administrative applications dominate the current landscape, where AI is concerned. Almost a third of respondents use AI for correspondence and reporting; nearly three in ten apply it to clinical documentation. ChatGPT is the clear favorite tool, used by 44.7% of respondents. Used effectively, these tools can dramatically cut the time you spend on paperwork – and free up hours for client support.

Higher-stakes domains are starting to take shape, too. In November 2025, a federal executive order encouraged the adoption of predictive analytics and AI in foster care. Then, in mid-2026, the Administration for Children and Families allocated $6 million toward testing such technologies in child welfare. Risk assessment and placement algorithms are the obvious examples, and errors can have lifelong repercussions for clients.

Why an Online MSW Is Worth Pursuing Now

The field offers attractive career prospects for anyone weighing an online master’s program in social work. The Bureau of Labor Statistics projects 6% employment growth for social workers between 2025 and 2035 – well ahead of the roughly 3.5% across all occupations. Meanwhile, the median salary in May 2025 was $61,780.

An MSW is also required for clinical licensure in all 50 states. Studying online lets you keep working while you qualify, thanks to the asynchronous nature of many classes.

Then there’s the fieldwork. CSWE-accredited MSW programs include at least 900 hours of it, and these can typically be completed locally. That’s where you may well encounter AI-assisted documentation for the first time, with a field instructor on hand.

Furthermore, the Council on Social Work Education has made AI the lead focus of its 2026 to 2030 strategic plan. You’re likely to see more training in the domain as a result.

What Constitutes AI Literacy in Social Work?

Literacy, in this context, means being able to wield the technology effectively – while still applying a critical lens to its output. A 2025 article in the Journal of the Society for Social Work and Research goes further. It highlights recognizing potential bias in AI systems as a core competency.

Day to day, that looks like:

1. Double-checking any records produced by AI-assisted documentation tools
2. Being aware of how client data is likely to be used downstream
3. Questioning how a risk score was built before relying on it
4. Staying mindful of paraphrasing that distorts a client’s own words

This matters, because plenty of practitioners feel underprepared. Over a quarter of Moritz respondents cited lack of training as a challenge. Around three in seven said they had no say in which technologies their workplace adopted. Being able to critically evaluate each tool arms you to contribute to those discussions.

Navigating Ethical Dilemmas in AI-Assisted Social Work

Data privacy is the chief concern among social workers using AI, cited by 46.5% of Moritz respondents. And with good reason. The NASW Code of Ethics binds you to protect client confidentiality, but using an unapproved chatbot to process identifying client details can constitute a breach. Intent doesn’t change that, either – however tempting the shortcut feels at 6 p.m. on a Friday.

Informed consent and bias matter just as much. Clients should know when AI plays a role in their care (it might be something like having an AI transcription tool in the room). Ideally, they should also be able to request a human review. As for bias: many predictive tools learn from historical data. If those records reflect discriminatory attitudes, which, sadly, is quite possible, the algorithm can absorb them. It may then perpetuate those prejudices at scale.

Policymakers are taking note. Illinois and Nevada have already banned AI from delivering behavioral health care, and other states are moving to restrict AI-driven chatbots for therapy. Colorado’s 2026 law goes a step further. It limits how licensed social workers themselves can use AI.

The profession is responding, too. Two-thirds of Moritz respondents identified ethical guidance as their most pressing need regarding AI. Recognizing that, NASW’s Code of Ethics Revision Workgroup is now recommending supplemental guidance.

Developing Professional Competence and Ethical Judgment in AI Use

Establishing healthy boundaries around AI tech is now an important part of building professional competence, and will only become more so. It starts with studying the AI policy of your chosen MSW program. Asking your field instructor how their agency vets technology purchases is good practice, too.

Cultivating these habits early lets you bring fresh eyes to your agency’s challenges in adopting new technologies. Honestly, that’s a real asset.

Why? Because the profession is divided. In the Moritz survey, 44% of practitioners called AI a negative development, yet nearly three-quarters expect it to play a growing role. Hybrid approaches are likely to emerge, as organizations try to balance efficiency with their duty to avoid harming clients, through premature adoption. A new graduate who can explain those nuances to a concerned client (or indeed, supervisor) should be in high demand.