There is a debate regarding the ethics of using artificial intelligence in social work. Should practitioners embrace it, or should we avoid it altogether and not feed the beast? Let us explore its pros and cons through a social work lens with a macro perspective. According to the Grand Challenges for Social Work (2023), one of the grand challenges is to “Harness Technology for Social Good.” Traditionally, social workers have addressed complex social issues and gathered essential data to serve marginalized communities and vulnerable populations (Grand Challenges, 2018). However, usually, this involves a strenuous amount of time, causing social workers to wonder if or how there can be an easier way to bring about change. By not embracing technological advancements in their careers, social workers risk missing out on tools that could significantly enhance their social work practice. In an ever-evolving world, it’s imperative for the profession to stay updated and leverage the best tools available.With new advancements in technology, the profession of social work must adopt new technologies such as information and communication technologies to enhance the efficacy of social programs, drive social discovery, and transform the social work profession through advances in technology ethically to reduce inequalities (Grand Challenges, 2018). Thus, we must integrate technology into our field, fostering a more equitable profession for practitioners and consumers.
One way to revolutionize social work is by integrating artificial intelligence into our assessments. AI makes assessment implementation quicker and easier across micro, mezzo, and macro levels, aligning with the General Intervention Model (GIM). GIM is a method commonly used in social work practice that was introduced in the1960s by the Milford conference, which advocated that a generic approach to social work was necessary (Schatz et al., 1990). Thus, the GIM, promoted at CBU’s Division of Social Work, aids practitioners in reducing inequalities and driving community transformation. As we all know, engaging clients and assessing their strengths and needs is essential to our social work practice to enhance our communities. To start the process of enhancing communities and reducing inequalities, the GIM starts with engagement and follows with assessments (Angy, 2020). GIM allows social workers to examine the evidence with an assessment objectively. Therefore, regarding macro social work, social workers must conduct a community’s needs assessment to understand the strengths and needs of communities rather than making unfounded assumptions because every community has unique needs, strengths, cultures, worldviews, and assets (Rossi et al. (2019). One way to tackle community needs assessments is with the help of a new emerging technology from OpenAI (2023), ChatGPT. This blog is about my experience with using ChatGPT to help with the statistical analysis of the results of a community needs assessment I conducted in Moreno Valley, why it is essential to conduct a community needs assessment, the dos and don’ts of using ChatGPT, and to explain why I used it, how I used it, what I found and my recommendations.
What is a community needs assessment, and why is it important? A community needs assessment (CNA) is a tool that helps social workers understand the needs and characteristics of a community as well as their available resources for mobilization through data collection using quantitative and qualitative methods such as surveys and interviews (Rubin & Babbie, 2017; National et al., 2023; Ohmer & Underwood, 2013; Chow & Carol, 2013; Community Toolbox, 2023). CNAs help identify gaps in local services, prioritize issues, help with advocacy, obtain funding, and plan interventions (Community Commons, 2023; Taylor & Mackenzie, 2023; Community Toolbox, 2023). Rossi, Lipsey, & Henry (2019) defines a needs assessment as identifying and understanding a particular population’s needs with the goal of improving programs, services, or policies.
Overall, studies demonstrate the efficacy of community needs assessments in various fields using a mixed methods approach, leading to positive outcomes such as increased community participation, stakeholder commitment, and empowerment (Welch, 1988; Ravaghi et al., 2023; Community Commons, 2023; Taylor & Brunson, 2023; Community Toolbox, 2023). Consider this: Without the community’s perspective and involvement, how can we truly understand its needs and strengths to effect change? The answer is that we cannot without incorporating our own biases, which can lead to oppressive social work. In conclusion, CNAs are essential for social workers to foster community engagement and the knowledge needed to transform communities by focusing on the community’s perspective, not the social workers.
So, what is OpenAI (2023) Chat GPT? OpenAI’s ChatGPT (2023) is a revolutionary language generation model with advanced data analysis. It can understand and process unstructured data, create text resembling human writing, and work with various tools and technologies (Ahmad, 2023). There are two different versions of ChatGPT, ChatGPT-3.5 and ChatGPT-4 (updated version). Now you might ask, is it logical to use ChatGPT? The answer is yes, because scholars find it significant for analyzing large amounts of data (Ahmad, 2023). It can create reports, summarize large datasets, and has automation capabilities for repetitive tasks, such as data entry and cleaning, enabling social workers to focus on other essential research tasks (Chandra, 2023; Ahmad, 2023; Xin, 2023; Sheppard, 2023). Even though its capabilities are gaining recognition, a gap in the research exists, including a lack of studies to examine its significance in general and ethical concerns regarding confidentiality, such as client details (Maxwell, 2023). Therefore, I tested its capabilities to aid future research and address existing gaps.
Significantly, social workers must understand the limitations of using OpenAI (2023) ChatGPT version 3.5. First, it is essential to note that ChatGPT-3.5 relies on its training data, while GPT-4 can fetch real-time information with the right plugins. Therefore, ChatGPT-3.5 cannot provide information that emerged past September 2021. Anything social workers want to inquire about using ChatGPT 3.5 will be generalizations and guesses that sound logical, and it is not using the dataset in its training. Therefore, the first don’t for ChatGPT is not to use it to answer relevant questions, especially when the information can put people’s lives on the line or cause harm. As we know, social workers must use evidence-based practices, such as peer-reviewed research studies with significant results, to support our actions. Thus, Social workers must research ChatGPT’s information, even with GPT-4 and its plugins, to ensure that the information retrieved is evidence-based. In addition, OpenAI (2023) ChatGPT is a brilliant writer that can write papers about anything within minutes. However, let us suppose scholars use this capability to write their papers or scholarly articles. In that case, they risk getting into trouble for plagiarism and can get their degrees revoked due to artificial intelligence detection websites such as ZeroGPT. ZeroGPT is a website to detect if artificial intelligence is writing a paper or if AI created any text. Therefore, having AI write papers will be counted as plagiarism or academic dishonesty if detected.
creating a survey for a community needs assessment with GPT-3.5, it is good at creating surveys on the spot. However, it is unproductive to do so without GPT-4 because GPT-3.5 survey questions will not be evidence-based. As mentioned earlier, ChatGPT can only retrieve evidence-based information with the appropriate prompts and GPT-4 plugins. Therefore, when using ChatGPT-3.5 to create survey questions, it may create a question that sounds evidence-based but is a generalization. To do so, social workers must use Chat GPT-4, which takes a monthly subscription to get access. It provides access to plugins, allowing social workers to prompt ChatGPT-4 to open websites, open pdf files, and provide evidence-based questions from the community needs assessment survey from the article it finds. For example, social workers can prompt: “Find a relevant Community Needs Assessment research article, open the pdf, and provide me with its evidence-based survey questions.” Thus, the questions social workers create for their community needs assessment survey will need to be backed by valid and reliable studies, which ChatGPT cannot do unless prompted to do so with the proper plugins and monthly subscription, which will still need to be researched and studied.
Furthermore, OpenAI (2023) ChatGPT has a memory and word limit. When social workers have it analyze large amounts of statistical data for summarization or to create an outline, ChatGPT can start to forget the information social workers have entered after so many prompt entries. Imagen hours spent providing prompts to have it summarize data, and it ends up giving you made-up summaries that sound relevant because you gave it too many prompts? Over prompting is a mistake many make without knowledge of its word limitations. Therefore, social workers need to use it moderately or in chunks to ensure that ChatGPT is not making guesses or generalizations based on the statistical data entered. Lastly, ChatGPT 3.5 is incapable of providing accurate reference citations. Therefore, when social workers ask it to provide evidence-based citations, it will create made-up citations that sound logical based on generalization. To enhance social work practice, social workers must consider these limitations.
ChatGPT may also produce biases based on its programming, leading to outputs that do not align with social workers’ values and ethics. Also, ChatGPT lacks empathy and cultural competence because it can generate responses that may appear empathetic; however, it does not understand, nor can it relate to, human emotions and cultural perspectives. Therefore, social workers must use caution when using ChatGPT for engaging in sensitive topics or culturally diverse populations. Lastly, ChatGPT relies on clear and specific instructions to provide valuable outputs. When social workers feed ChatGPT vague or ambiguous prompts, it can lead to vague or unhelpful answers. Therefore, social workers must provide ChatGPT with clear instructions to avoid miscommunications or misunderstandings.
Despite OpenAI (2023) ChatGPT limitations, it has many benefits that can enhance social work practice. First, ChatGPT can accurately analyze large quantities of data to provide social workers with a descriptive analysis by summarizing, organizing, and outlining the information. Moreover, it can extract the main points and themes from large data sets to offer valuable insights for an exhaustive descriptive analysis. Also, ChatGPT can help with statistical analysis by providing the sample size, standard deviation, standard of error of the mean, margin of error, confidence interval, and power analysis. Second, ChatGPT can create consent forms, confidentiality agreements, commitment agreements, and other essential forms social workers use within minutes. Third, social workers can utilize ChatGPT to proofread their work, offer suggestions to improve it, and provide insights on needed information. For example, when writing a grant, a social worker can copy and paste a part of a grant section’s requirements and then copy and paste what they wrote to see if their work aligns with the grant to ensure that they are addressing everything that the sections of the grant are requiring. Fourth, OpenAI (2023) has developed a new edition called ChatGPT 4 that can assist social workers in brainstorming ideas and creating potential solutions for complex social issues. In addition, it can help social workers by providing summaries of research articles when they copy and paste the articles into sections or use the plugins to open pdfs for it to read when providing a link to it. However, ChatGPT 4 still has memory limitations despite the expanded memory and plugins, like ChatGPT 3.5. Overall, OpenAI (2023) ChatGPT saves social workers time and money and provides valuable insight that can improve the social worker’s quality of work. Therefore, ChatGPT is essential for addressing the 7th Grand Challenge of Social Work (2023) to “harness technology for Social Good.”
Significantly, I used OpenAI (2023) ChatGPT because I needed to be able to analyze 319 qualitative questions for a community needs assessment I conducted for Moreno Valley. I created a mixed methods survey for my community needs assessment with eleven quantitative and three qualitative questions with 125 participants. I was overwhelmed and had no idea how to find the time to analyze the qualitative answers, which were crucial to understanding the community’s perspectives on Moreno Valley’s strengths and needs, especially with no funds. Therefore, I decided to copy and paste the answers, one question at a time to have ChatGPT summarize, outline, and create themes of the answers to determine the consensus regarding the most pressing needs of Moreno Valley and its strengths to identify its assets. Since I conducted the survey using Google Forms, it automatically organized the answers to the quantitative/qualitative questions into an Excel sheet. Also, it created charts to be able to study the answers effectively.
Regarding the qualitative answers, once I had ChatGPT summarize the qualitative answers, I realized how important and powerful it was. Thus, I decided to statistically analyze all the answers to the survey to produce statistical data such as the margin of error, standard deviation, mean, and confidence levels within minutes. Overall, OpenAI (2023) ChatGPT is the best tool for social workers to use to save time and money because we can analyze the data for free without using statistical software such as Stata. Therefore, I recommend that all social workers use OpenAI (2023) ChatGPT to save costs and analyze their data for descriptive statistical analysis. Also, I recommend that social workers use this technology to create forms and proofread their work when writing grants to ensure that their grants cover all their requirements. Social workers need to embrace artificial intelligence, instead of ignoring or being afraid of it, to utilize it for the greater good because social workers are obligated to be competent in their social work practice. We have enough on our plate with all our communities’ challenges and inequalities, and saving time and money is crucial to tackling those challenges and inequalities head-on. Therefore, a call to action is needed for all social workers to experiment with ChatGPT and keep its limitations and strengths in mind to help advance the social work practice to create equitable environments.
References
Admin. (2022, December 30). Chat GPT: How ai will and can change your life. for the better – for now. Siesta Codes. Retrieved April 15, 2023, from https://siesta.codes/en/chat-gpt/
Ahmad, R. (n.d.). “unlocking the potential of data analysis with ChatGPT”. LinkedIn. Retrieved March 26, 2023, from https://www.linkedin.com/pulse/unlocking-potential-data-analysis-chatgpt-romeer-ahmad.
Angy. (2020, December 30). Social Workers and the generalist intervention model: Social Work Haven. Social Work Haven | Become A Social Worker. Become A Better You. Retrieved April 2, 2023, from https://socialworkhaven.com/generalist-intervention-model/
Brunson, T. T. andM. (n.d.). Community needs assessments. Community Needs Assessments | Learning to Give. Retrieved April 3, 2023, from https://www.learningtogive.org/resources/community-needs-assessments#:~:text=Community%20Needs%20Assessments%20seek%20to%20gather%20accurate%20information,assessments%20establish%20the%20essential%20foundation%20for%20vital%20planning.
Chandra, N. S. (n.d.). How to use CHATGPT to enhance data analysis workflow. Kanaries. Retrieved March 26, 2023, from https://docs.kanaries.net/articles/chatgpt-data-analysis.en
Chatgpt for analytics – 6 compelling use cases. Graphable. (2023, March 21). Retrieved March 26, 2023, from https://www.graphable.ai/blog/chatgpt-for-analytics/
Chow, Julian Chun-Chung , & Peng, Carol (2013). Community-Needs Assessment. obo in Social Work. doi: 10.1093/obo/9780195389678-0059
Community needs assessments: Mapping and data resources: Research: National recreation and park association: NRPA.National Recreation and Park Association. (n.d.). Retrieved April 2, 2023, from https://www.nrpa.org/publications-research/evaluation-resource-hub/community-needs-assessments/
GPT-4 and CHATGPT detector by zerogpt: Detect openai text. ZeroGPT. (n.d.). Retrieved April 15, 2023, from https://www.zerogpt.com/
Grand Challenges for Social Work. (2023, April 14). Retrieved April 15, 2023, from https://grandchallengesforsocialwork.org/#the-challenges
Harness technology for social good – grand challenges for social work. (n.d.). Retrieved April 16, 2023, from https://grandchallengesforsocialwork.org/wp-content/uploads/2015/12/180604-GC-technology.pdf
Maxwell, T. (2023, January 6). 9 questions about CHATGPT, answered. MUO. Retrieved March 26, 2023, from https://www.makeuseof.com/chatgpt-questions-and-answers/
Nurwisah, I. (2014, January 23). Life is beautiful? even the marginalized can benefit from art. HuffPost. Retrieved April 15, 2023, from https://www.huffpost.com/archive/ca/entry/life-is-beautiful-even-the-marginalized-can-benefit-from-art_b_4172371
Ohmer, M., & Underwood, E. (2013). Community needs assessment. In Encyclopedia of Social Work. Retrieved 10 Mar. 2023, from https://oxfordre.com/socialwork/view/10.1093/acrefore/9780199975839.001.0001/acrefore-9780199975839-e-73.
OpenAI. (2021). ChatGPT. https://www.openai.com
Team, P. M. Y. (2023, April 11). How to use chat GPT, complete guide, Create Account & Login. PmmodiYojana.in. Retrieved April 15, 2023, from https://pmmodiyojana.in/how-to-use-chat-gpt/
Schatz, M. S., Jenkins, L. E., & Sheafor, B. W. (1990). Milford redefined: A model of initial and advanced generalist social work. Journal of Social Work Education, 26(3), 217-231.
Section 13. conducting surveys. Chapter 3. Assessing Community Needs and Resources | Section 13. Conducting Surveys | Main Section | Community Tool Box. (n.d.). RetrievedApril 2, 2023, from https://ctb.ku.edu/en/table-of-contents/assessment/assessing-community-needs-and-resources/conduct-surveys/main
Section 1. developing a plan for assessing local needs and resources. Chapter 3. Assessing Community Needs and Resources |Section 1. Developing a Plan for Assessing Local Needs and Resources | Main Section | Community Tool Box. (n.d.). Retrieved April 3, 2023, from https://ctb.ku.edu/en/table-of-contents/assessment/assessing-community-needs-and-resources/develop-a-plan/main
Study on CHATGPT BY CHATGPT: The Survey Design Experience and Results Analysis. Milieu Insights. (n.d.). Retrieved March 26, 2023, from https://www.mili.eu/insights/study-on-chatgpt-by-chatgpt-the-survey-design-experience-and-results-analysis