The emerging use of Artificial Intelligence (AI) in social work

Introduction

By Social Work England

AI systems are developing rapidly, and their use is becoming increasingly common. Changes in social work practice are inevitable as more social workers, students and academics use AI. We therefore need to understand the opportunities and challenges of AI in social work, so we can continue to:

• promote and maintain professional standards for social workers in England
• protect, promote and maintain the health, safety and wellbeing of the public
• promote and maintain public confidence in social workers in England 

The research

In 2025 we commissioned 2 pieces of research that explore the emerging use of AI in social work. Both pieces of research took place in spring 2025. They explored:

• The key messages from reliable, high-quality research on AI, with focus on social work and closely related sectors.
• The types of AI being used across health and social care in England and their application in social work practice, including the risks of bias and discrimination.
• Issues of data protection and confidentiality when using AI with people using services and the public.
• To what extent social workers felt confident and prepared to use AI ethically and appropriately and in line with Social Work England’s professional standards.
• How employers were supporting social workers to use AI ethically and appropriately.
• The areas of Social Work England’s professional standards which may be impacted by social worker’s use of AI in their work.
• How social work education providers were preparing students for AI in their future work.

Literature review

We commissioned The Open University (OU) to carry out a literature review using a Rapid Evidence Assessment of the Literature (REAL) approach. The study looked at UK and international literature relating to the use of AI in social work, health and social care. This included peer reviewed journal articles as well as a range of grey literature [1].

Its findings explored implications for the use of AI both in social work practice and social work education. A panel of people with lived experience and a panel of social workers were involved in the study to ensure its relevance to current social work practice. 

The OU reviewed and analysed a total of 119 full text articles, and 44 pieces of grey literature.

• View the full research report by The Open University

[note 1] Grey literature is information not produced by commercial publishers and is not published in scholarly journals. Grey literature may include multimedia content such as blogs and web pages.

Research investigation

We also commissioned Research in Practice to investigate how AI is already being utilised within social work and social work education, and how AI in social work might develop in the future. It used a mixed methods approach which included surveys, interviews and focus groups to engage with the following stakeholders:

• practicing social workers
• social work employers
• social work educators
• people with lived experience of social work

Research in Practice received 203 responses to their survey, 155 of whom were social workers.

• View the full research report by Research in Practice

Summary of key findings

This summary sets out key findings identified across both pieces of research. The full reports for each piece of research can be accessed using the links at the top of this webpage. Here you can access full details of the methods used, findings, analysis, discussion and suggestions for further action.

Use of AI in social work settings

Findings from both pieces of research showed that there were both benefits and challenges to using AI in social work. Benefits included improvements in efficiency and accessibility and reduction of costs. Challenges focused on ethical considerations and specific challenges around the integration and implementation of AI in social work settings. Both reports found that generative AI (‘GenAI’) is the most used type of AI in social work settings. GenAI creates original content, including text or images, based on patterns in existing data. 

The most common AI tools currently used in social work were found to be:

• AI virtual assistants
• transcription software
• case recording support (generated documentation)
• chatbots

The research investigation gathered feedback on social workers’ perceptions of AI use in social work, as well as their experience of using (or not using) AI in their practice. When asked whether they used AI as part of their practice, of the 155 social workers who completed the survey:

• 40% said they have used AI with direction from their employer.
• 24% said they have used GenAI without direction from their employer.

The rollout of AI across the social work profession was found to be uneven. AI was commonly used in some settings and in others there was little awareness of AI. Social work employers were noted to have taken varying approaches to determining acceptable use of AI in settings where this was already in use. For instance, while respondents indicated that some employers prohibited the use of AI during face-to-face interactions, others promoted the use of AI in home visits. In some cases, social work employers had procured bespoke AI products to support note taking and the creation of case notes. Whereas others had made use of publicly available AI to support administrative tasks.

Efficiency, accessibility and reduction of costs

The literature review identified some key potential advantages of rolling out AI in social work. Their findings suggested that AI may be able to:

• increase access to services
• support decision-making
• support risk assessments
• increase workflow efficiency
• increase social workers' efficiency
• reduce workload
• enhance collaboration
• provide personalised and tailored support
• improve service quality

The research investigation survey found that respondents were positive about the overall benefits of AI in social work:

• 83% of respondents felt ‘AI had [the] potential to reduce administrative burden for social workers.
• 70% of respondents indicated that there are ‘equal or more benefits to using AI in social work than there are risks.
• Respondents were positive about the potential for AI to facilitate workload management and administrative tasks. However, respondents were less optimistic around AI’s ‘capacity to aid decision making’ (48%) or to ‘identify risk and need in social work’ (46%).
• Respondents suggested that AI tools have the potential to improve the quality and consistency of case recordings and communication.
• Some respondents indicated that time saved on administrative tasks allowed for a greater focus on face-to-face engagement with clients.

Benefits around workforce wellbeing were also identified as a key theme in responses, with some respondents linking this to reduced cognitive load and stress for social workers. This may have broader positive implications for the recruitment and retention of social workers.

Similarly, the research investigation found in some scenarios AI assistive technology can help bridge learning and language gaps. This improved accessibility between social workers and the people they work with. It may also act as an assistive technology for many social workers who are neurodiverse.

Risks and ethical concerns

While the use of AI in social work was associated with potential benefits, both studies identified a range of risks and ethical concerns. 

Privacy and data protection concerns were key themes in the findings from both reports. For instance, some participants in the research investigation focus groups noted examples where personal information was shared with publicly available GenAI such as ChatGPT, Claude or Microsoft Copilot. Both reports note that the ways in which these systems process or store data is often unclear. The research investigation highlighted the need to raise awareness of potential risks and issues related to privacy and data protection with social workers.

A particular challenge identified was whether there was a need for informed consent when using AI in social work. In the research investigation social workers discussed the complexities around capturing informed consent when using AI. Consent to record a conversation or meeting for transcription was considered as relatively straightforward but became more complex when working with those using social work services who might be vulnerable or lack capacity.

Concerns around fairness, bias and discrimination were also frequent themes in both pieces of research. The literature review suggested that AI systems could amplify existing biases in data, and the risk that this may result in unfair outcomes for some (often marginalised) demographics.

Both studies also raised concerns regarding the accuracy and reliability of AI generated text, including case notes. This is of particular importance where AI is involved in shaping or recommending case outcomes. Given the range of challenges highlighted in both studies, accountability and responsibility were seen to be of particular importance. However, where AI was in use in social work settings, it is not always clear who is accountable for the outputs of AI and whether this responsibility sits with the social workers, their line manager or those who created the AI software. Comments from a range of respondents, as well as the literature, emphasised that, social workers should be responsible for the accuracy and reliability of their case recordings and how they use AI.

Employers also have a key role to play in responsibly procuring AI tools, preparing and enabling social workers to use AI tools, and ensuring rigorous policies are in place to guide the use of AI. This includes developing effective methods of quality assuring AI outputs and ensuring that use of AI is supervised appropriately in line with professional standards.

Social work training and education

It is important that social work students and apprentices develop up to date digital literacy skills and receive adequate support to use AI lawfully, ethically, and responsibly in social work practice, and to be able to use this to support their learning.

The need to equip social work students and apprentices for a fast-changing professional context should be a key consideration in social work education and training. However, survey responses indicated 86% of social workers that graduated in the past 5 years did not receive any specific preparation on using AI in social work practice during their education and training.

As with findings regarding social work employers, engagement with social work educators showed variance in the extent of their engagement with AI. The literature review identified a range of potential benefits of using AI in social work education. These included:

• the potential for personalised learning, access and support
• improved administrative efficiency
• equipping students with digital literacy required for future workplaces
• enabling assessment and curriculum redesign

These potential benefits were mirrored in findings from the research investigation. However, this also heard that AI is often considered in relation to academic integrity, rather than as a tool which might support and enhance learning. Social work educators participating in the research investigation stated that attitudes and approaches towards using AI in social work education vary widely.

The investigation emphasised that guidelines around the use of AI in social work education are often unclear and greater support for social work educators in this area would be beneficial. 

Gaps in available research

The literature review included a broad range of both academic and grey literature. This accessed studies on AI in health, social care and social work both internationally and in the UK. The researchers noted a lack of empirical evidence [2] around the use of AI in social work. While the review did identify some empirical research, it largely utilised grey literature. Most available literature covered AI in health and social care, without a specific focus on social work. However, considerations across the health and social care sector were often parallel to those in the social work context. Given the emerging nature of AI technology further research will be essential for us to continue to understand its evolving use in social work. This will ensure its use leads to positive outcomes for social workers and the people they work with.

[note 2] Empirical evidence comes from real world experiments, observations and experience rather than from ideas or theories.

Recommendations

Research colleagues from The Open University identified several key considerations which may support social workers as they use AI in their practice, including:

• An ongoing focus on ethical practice, with an awareness of rapidly evolving social work practice.
• The importance of governance and regulation of AI systems, both at the level of the employer, and at a national level.
• The importance of putting people at the heart of social work relationships and interventions.
• An emphasis on critical thinking, professional judgement and decision making.
• The value of continuing professional development (CPD) where practice and associated AI tools are rapidly evolving.

Research in Practice also highlighted broad policy areas where change may support social workers to use AI in a lawful, ethical and responsible manner, including:

• Government departments, regulators, professional bodies and other organisations with responsibility for social work should work together to inform and shape the use of AI by social workers. This work should build on existing standards and frameworks .
• Social Work England should consider where updates may be necessary to their ex-isting standards and guidance, including the professional standards, to reflect in-creasing use of AI and how this may impact social work practice. As part of this work Social Work England should consult with people with lived and learned experience of social work.
• Social Work England should continue to work with partner agencies to better understand the skills, knowledge and behaviours social workers need to use AI law-fully, ethically and responsibly.
• Social work education providers should ensure that they are supporting social work students and apprentices to use AI responsibly both during their course and in their preparation for practice. 

What happens next?

The findings from the research have been useful in improving our understanding of the current use of AI in social work education and practice. We appreciate that this is a rapidly evolving area which will continue to pose questions, challenges and opportunities for social work employers, educators and practitioners over the coming years.

Social Work England will carefully consider the messages from these reports. We endorse the recommendations made by Research in Practice and The Open University and will determine how we might best respond to any areas that link directly with our role as a regulator.

We are conscious that many of the findings  and recommendations here encompass issues that involve a wide range of groups. Collective leadership will be needed from many organisations including professional bodies, systems regulators and employers to help embed the safe and ethical use of AI in social work. We will therefore work with the sector to ensure that we are able to take learning from these 2 pieces of research to support positive developments around AI in social work practice and education.

In collaboration with Skills for Care, we are convening a free one-day summit focused on AI and the future of social work in Birmingham on Wednesday 20 May 2026. This event will bring together colleagues from across social work, people with lived experience, policy leaders, technologists and academics.