From recipe suggestions to emails to unofficial mental health diagnoses, artificial intelligence is already integrated into the everyday lives of many of the clients social workers support.
The pace of change can be dizzying: how can social workers keep up with this evolving technology, and assess how ChatGPT, Replika, Claude and countless other AI applications are impacting their clients and their own practices?
In 2026 and beyond, OASW is exploring the applications and impact of AI on the social work profession. This spring, the association hosted two webinars delivered by practitioners with extensive experience using artificial intelligence and educating clients and other mental health practitioners on its uses.
It's important to note that AI is a massive topic, and there is no simple or singular way to interpret the impact of AI on social work, or the individuals and communities social workers support.
AI, Chatbots, and Your Future Social Work Practice: Navigating Risks and Ethics in Therapy
An April webinar from Lawrence Murphy explored the ethical concerns social workers face when AI tools aren't aligned with social work values, and how AI is shaping the future of therapy. Murphy is an instructor at Wilfrid Laurier University's Department of Psychology and the founder of Worldwide Therapy Online, the world's first online clinical practice, established in 1994.
"The challenge with AI, with these chatbots, is that they have no morals, they have no ethical principles, they have no best practice fundamentals," he told attendees.
A major theme was dependency. Large Language Models (LLMs) are tuned to exploit attachment, suggested Murphy.
People with attachment issues and those who experience intense emotional dependency are commonly people who seek counselling and are at risk of becoming dependent on LLMs.
He referenced a study which suggests there may be a dependency process at work underlying the ways in which people are using chatbots. While only 12% of respondents said they were using the chatbot for companionship, 51% used terms like friend, companion or romantic partner when describing the relationship, and 80% featured seeking emotional and social support as the main theme of chats.
Lawrence Murphy's June 10th workshop expanded on the themes discussed in this webinar, and also explored how the underlying construction of AI systems can impact individuals seeking care and complicate our perceptions of morality and ethics in care.
To contact Lawrence Murphy, email him at: [email protected]
From Automation to Advocacy: Navigating AI with Social Work Values
Dr. Badillo-Diaz, who divides her time between New York and Toronto and teaches AI courses for social work practice at NYU, Columbia, and the University of Tennessee, said AI has a practical role to play in how social workers build resources, develop programming, and support their clients.
In her presentation in March, as part of OASW's Social Work Week and National Social Work Month learning series, she helped social workers develop a practical understanding of the ways AI can help streamline aspects of their practice and potential risk areas.
She stressed that while AI is capable as a notetaker, resource gatherer and programming assistant, practitioners must be aware of the accompanying risks. AI, Dr. Badillo-Diaz said, is not:
- Free from biases or inaccuracies;
- Secure with data and privacy;
- or a replacement for professional expertise.
"The human must remain in the loop, we must protect, right, because we don't want high-stakes decisions impacting our most marginalized and most vulnerable," she said.
To guide social workers' use of AI, Dr. Badillo-Diaz suggests the "COSTAR" method, an acronym for the words Context, Objective, Style, Task, Actor, and Result.
This approach maximizes the quality of AI-generated outputs by providing context for the task, giving specific instructions, and refining the output through follow-up prompts. This applies to creating materials for clients or curriculum content, for example.
"We want to preserve our professional judgment and not just rely on the decisions and outputs that AI has given," Dr. Badillo-Diaz said.
Dr. Badillo-Diaz produces tools and training for social workers. Visit The AI Social Worker to learn more and access resources specifically created for social workers navigating A.I.
To learn more about evolving challenges and themes in social work from experienced practitioners, check out our upcoming OASW Learning Centre events.