Artificial Intelligence in Social Work: Emerging Ethical Issues
The ethical tension in Reamer’s article that makes me most uncomfortable is the idea that AI could start shaping the tone or direction of a client’s experience without me realizing it. I am used to thinking of building a relationship as something that it is created through intentional, physical human communication with empathy and honest judgment. Reamer’s point that AI tools can respond in ways that feel overly personal, or overly formal is something that stuck with me because for me, a person’s natural tone is how you build trust. I have already seen a small version of this tension in my own work during undergrad. At the free clinic that I interned at, some clients had to rely on automated messages or digital systems that were meant to guide them through their medical applications. I noticed how those messages, most of the time, confused them. I remember the hospital using automated messages in English to provide guidance, and Spanish speaking clients feeling defeated because they could not understand. One client told me, “They don’t make this easy for us, like they don’t want us to get help.” This stuck with me, because AI was designed to ease our concerns, and instead there are many times where it just complicated things.
I agree with Reamer that AI can create ethical risks we don’t always see coming, especially when it shifts the tone of communication or replaces human judgment. But I also pish back on the idea that AI is completely harmful. I think it will depend on how we use it and how closely we monitor its impact. Some AI tools can really help clients access information more easily. If a client were to ask me about this tension today, I would say something like “Technology can be helpful, but it can also confuse us or feel impersonal. It is my job to make sure any tool we use supports you, and not make you feel like it is replacing our relationship. If it feels wrong to you at any point, we can talk about it and adjust together.”
Haidt in the Room with Your Clients
Even though Haidt and Rausch’s evidence focused on adolescents, I kept thinking about my experience with the older adults I worked with during my undergrad capstone project. Their first line of evidence comes from young people’s reports about feeling more anxious or lonely after using social media, and I have seen similar reactions in older adults (Haidt and Rausch, 2026). This evidence made me wonder how these same patterns show up later in life. Especially for older adults who are already struggling with isolation and limited social support. In the TED talk, Haidt’s point that humans are “ultrasocial” stuck with me because I have seen how deeply older adults crave connection, and how technology can either support that craving or complicate it (Haidt, 2026, 1:15).
Most of the older adults I interacted with relied on Facebook. It was their main source of connection to their families, especially grandchildren. I also saw how using Facebook and other messaging platforms sometimes became overwhelming for them. Some clients didn’t know how to use all the different features to not only Facebook, but their smartphones. Haidt’s argument about how technology interferes with how humans are wired for connection made sense to me here when thinking back to my own experiences. Even though older adults aren’t adolescents, they still feel the emotional effects from how social media influences their connections. I have watched some clients become more anxious or withdrawn after spending too much time online, even though they had originally turned to digital platforms to feel less alone.
Although this can be the case sometimes, it is not always like this and that’s why I think Haidt’s argument doesn’t fully translate to this population. Many older adults use social media more intentionally, and it can become a tool that fills in the gaps to the barriers that they face like transportation, mobility issues, or the loss of a loved one. For some, social media doesn’t create harm and instead becomes a place where they feel seen. Overall, Haidt’s framework helped me name patterns I have witnessed, but also reminded me that technology’s impact depends heavily on context.
References
Haidt, J. (2026). How screens stole childhood — and how to get it back. Ted.com; TED Talks. https://www.ted.com/talks/jonathan_haidt_how_screens_stole_childhood_and_how_to_get_it_back
Haidt, J., & Rausch, Z. (2026). The case against social media: Seven lines of evidence. After Babel. https://substack.com/@jonathanhaidt/note/p-193691460
Reamer, F. G. (2023). Artificial Intelligence in Social Work: Emerging Ethical Issues. International Journal of Social Work Values and Ethics, 20(2), 52-71. https://doi.org/10.55521/10-020-205
Your reflections highlight something really important: both AI and social media can technically increase access, yet still create emotional or relational barriers depending on the person using them. What stood out most is how you’ve seen older adults experience the same loneliness and confusion Haidt describes in adolescents, which shows his framework has broader relevance than he claims. At the same time, you name a crucial nuance for many older adults, platforms like Facebook aren’t stealing connection but restoring it in ways their circumstances don’t allow offline. That tension you describe is exactly where practice gets complicated: technology can either deepen isolation or soften it, depending on context, literacy, and support. Your examples make Haidt’s argument feel less universal and more like a starting point that needs tailoring for each population.
Hey Kailey,
One issue I’ve come across often with older populations and even personally is the difficulty in telling if something that is come across online is AI or not. AI has become so sophisticated that many images and videos are difficult to distinguish from reality. My concern is that this can be used maliciously to warp and confuse one’s sense of reality. I think this is particularly likely for individuals who become trapped in echo-chambers. Although that is a separate consideration of AI, I think it’s worth speaking about, especially for people that are experiencing social isolation and find most of their community in online spaces.
I do appreciate that your brought up the issues of AI in automating certain services, and how that can make client’s feel left out or cause additional hurdle to their ability to gain access to resources. It can be frustrating having to speak to an automated AI system, rather than a human. I imagine this is especially the case when you’re in need and reaching out for help. I wonder how many clients gave up before being in a position to express these frustrations.
I do agree that intentionality with social media plays a huge factor in how it impacts the user – however, the nature of the algorithm is intentionally designed to keep users coming back, and it can be difficult for new users to establish guardrails in the settings to keep the notifications and pulls for attention to a minimum. From my own experience, many people that start off with the intention for specific usage end up slipping into the cycle of scrolling and being at the whim of push notifications in influence cycles.
Hey Kailey!
Thanks for your perspective. What stood out to me is that you and I did similar things in our Part 2 sections. You extended Haidt’s evidence to older adults, and I extended it to adults in recovery from stimulant addiction. In both cases, his framework partially fit and partially didn’t, and I think that is from the limits of building evidence on only one population and then generalizing from there. Your point about older adults using social media more intentionally made me think about the older people in my life and how they use social media. My thoughts went to my dad. He got a smart phone later in life, about seven years ago, at age 70, after using a flip phone forever. He uses Facebook once a day in the morning, only looks at family posts, never posts anything himself, and is done. He doesn’t “like” anything and never gives the algorithm a chance to learn about him. I wonder if part of what protects some older adults is that they came to these platforms as adults with fully formed brains and attention systems and they don’t feed the platform enough engagement data for the algorithm to hook them into an addictive cycle. I also noticed that some of your clients started out using Facebook intentionally to feel less alone and then became more anxious and withdrawn from spending too much time on it. I can relate to this. I first made a MySpace account in high school, but I never checked or posted on it. I don’t think the MySpace algorithm was built for harm the way Facebook’s eventually became. I got Facebook as a college freshman in 2004, and it took some years for social media to become something that hurt me. The harm was slow and insidious. Part of what made it stick, I think, was that I got on social media as a late teenager, when peer pressure to post and get likes was already shaping how I used it. Adults who came to the platform later, like my dad, never faced that same social pressure, so they were able to treat it as a tool instead of a stage. I did not see the detrimental effects of social media until I was already hooked. By then, the media was starting to become more and more aware of the negative aspects of social media. I think my experience supports your closing point about context. Haidt treats platforms as harmful in themselves, but whether they become a tool or a stage seems to depend a lot on who you were when you got there.
Hey Kailey,
I enjoyed reading about your perspectives on how AI can take the personal tone and trust from communication. In my practicum experiences, the agencies did not utilize AI for communication with clients, but I have seen this become an issue between professionals. In my opinion, professionals who use AI to rewrite their emails and networking communications can cause tension and a lack of trust between multidisciplinary relationships. Although this can seem dramatic, I have personally overheard conversations around the topic. For example, our organization received an email from a community partner who serves domestic violence survivors. It was obvious to us that AI was used to rewrite the inquiry due to the cold tone and directness of the message. Because it was in such contrast to previous communication, our Director was slightly offended by it and felt that it was not sincere due to the use of AI. While many individuals may disagree with his opinion on the email, it is important to note this experience for our use of AI writing tools when communicating with fellow professionals. In our work, relationships are important, and while individuals may use AI to sound more formal or professional, it can also cause harm to vital partnerships. This connects with Reamer’s warnings that good intentions can still lead to negative outcomes.
Your ability to apply Haidt’s and Rausch’s lines of evidence to older adults was admirable. While it makes sense that older adults would express negative effects of social media relatable to adolescents, I have had interactions with the population that contradict your experiences. In my discussion with older adults, they express their gratitude for the same social media platforms you mentioned as a way to stay connected and have a break from their responsibilities. It makes me ponder the various factors that contribute to their opinions related to whether or not social media can be harmful, beneficial, or a mix of both. As you mentioned, context is essential to better understand the perspectives of others. Meeting clients where they are is a key step in our helping process, so gaining your perspective through your own experiences has challenged my assumptions. Thank you for sharing your viewpoints so precisely!