#ai-ethics — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #ai-ethics, aggregated by home.social.
-
Digital ethics: it’s about more than just algorithms. It’s about ensuring fairness, transparency, and accountability in the data we use. Read the article in the comments
#digitalethics #techethics #ethicalai #AIethics #ai #technology #policy #fiarness
-
Digital ethics: it’s about more than just algorithms. It’s about ensuring fairness, transparency, and accountability in the data we use. Read the article in the comments
#digitalethics #techethics #ethicalai #AIethics #ai #technology #policy #fiarness
-
DATE: July 22, 2026 at 01:58PM
SOURCE: SOCIALPSYCHOLOGY.ORGTITLE: Seeking Empathy in the Age of Artificial Intelligence
Source: Association for Psychological Science
Artificial intelligence was initially imagined as a tool for efficiency. It could search, calculate, and write faster than humans, automating the tedious parts of work. But people are increasingly turning to AI for something more intimate: a shoulder to lean on. Recent surveys suggest that nearly a quarter of American adults and 13% of adolescents have used AI for socioemotional support or mental health advice. But can a machine provide empathy?
-------------------------------------------------
Private, vetted email list for mental health professionals: https://www.clinicians-exchange.org
Unofficial Psychology Today Xitter to toot feed at Psych Today Unofficial Bot @PTUnofficialBot
-------------------------------------------------
#psychology #counseling #socialwork #psychotherapy @psychotherapist @psychotherapists @psychology @socialpsych @socialwork @psychiatry #mentalhealth #psychiatry #healthcare #depression #psychotherapist #EmpathyInAI #AIMentalHealth #ArtificialIntelligence #TechAndEmotion #SocioemotionalSupport #AIethics #HumanVsMachine #PsychScience #MentalHealthTech #AICompanions
-
DATE: July 22, 2026 at 01:58PM
SOURCE: SOCIALPSYCHOLOGY.ORGTITLE: Seeking Empathy in the Age of Artificial Intelligence
Source: Association for Psychological Science
Artificial intelligence was initially imagined as a tool for efficiency. It could search, calculate, and write faster than humans, automating the tedious parts of work. But people are increasingly turning to AI for something more intimate: a shoulder to lean on. Recent surveys suggest that nearly a quarter of American adults and 13% of adolescents have used AI for socioemotional support or mental health advice. But can a machine provide empathy?
-------------------------------------------------
Private, vetted email list for mental health professionals: https://www.clinicians-exchange.org
Unofficial Psychology Today Xitter to toot feed at Psych Today Unofficial Bot @PTUnofficialBot
-------------------------------------------------
#psychology #counseling #socialwork #psychotherapy @psychotherapist @psychotherapists @psychology @socialpsych @socialwork @psychiatry #mentalhealth #psychiatry #healthcare #depression #psychotherapist #EmpathyInAI #AIMentalHealth #ArtificialIntelligence #TechAndEmotion #SocioemotionalSupport #AIethics #HumanVsMachine #PsychScience #MentalHealthTech #AICompanions
-
#TalDíaComoHoy nos dejaba #JoanFeynman. «Las mujeres no pueden hacer ciencia porque sus cerebros no están preparados” le dijo su madre cuando tenía 8 años. Eso no le frenó y le debemos la comprensión del origen de las auroras.
#machismoEstructural
#WomenInSTEM
@loretahur
www.loretahur.net
Ingeniera responsable de la Comisión IA de la #UniversidaddeDeusto #aiethicshttps://bsky.app/profile/loretahur.bsky.social/post/3mr7ipuhdx22m
-
#TalDíaComoHoy nos dejaba #JoanFeynman. «Las mujeres no pueden hacer ciencia porque sus cerebros no están preparados” le dijo su madre cuando tenía 8 años. Eso no le frenó y le debemos la comprensión del origen de las auroras.
#machismoEstructural
#WomenInSTEM
@loretahur
www.loretahur.net
Ingeniera responsable de la Comisión IA de la #UniversidaddeDeusto #aiethicshttps://bsky.app/profile/loretahur.bsky.social/post/3mr7ipuhdx22m
-
AI company plan to use irrigation water disappears after ABC question
By Ashleigh BarracloughData centre builder Firmus Technologies posts to its website a statement, since removed, indicating its intention to get some of its cooling water from the state-owned irrigation system, adding "but we do not want to highlight this yet, as they are yet to agree".
#AI #WaterResources #AIEthics #Utilities #StateandTerritoryGovernment #AshleighBarraclough
-
AI company plan to use irrigation water disappears after ABC question
By Ashleigh BarracloughData centre builder Firmus Technologies posts to its website a statement, since removed, indicating its intention to get some of its cooling water from the state-owned irrigation system, adding "but we do not want to highlight this yet, as they are yet to agree".
#AI #WaterResources #AIEthics #Utilities #StateandTerritoryGovernment #AshleighBarraclough
-
FYI: EDPB presses EU for data-sharing law as AI complaints strain regulators: Watchdogs want to exchange confidential files as AI complaints strain enforcement capacity, leaving the Commission to decide whether to draft the legal basis. https://ppc.land/edpb-presses-eu-for-data-sharing-law-as-ai-complaints-strain-regulators/ #DataProtection #AIRegulation #EDPB #DataSharing #AIEthics
-
FYI: EDPB presses EU for data-sharing law as AI complaints strain regulators: Watchdogs want to exchange confidential files as AI complaints strain enforcement capacity, leaving the Commission to decide whether to draft the legal basis. https://ppc.land/edpb-presses-eu-for-data-sharing-law-as-ai-complaints-strain-regulators/ #DataProtection #AIRegulation #EDPB #DataSharing #AIEthics
-
https://winbuzzer.com/2026/07/21/san-francisco-demands-apple-google-remove-13-nudify-apps-xcxwbn/
San Francisco has requested Apple and Google remove 13 nudification apps that create nonconsensual nude images.
#AI #Deepfakes #SanFrancisco #Apple #Google #GeneAI #AIImageGeneration #AIImageEditing #AIImages #AIEthics
-
https://winbuzzer.com/2026/07/21/san-francisco-demands-apple-google-remove-13-nudify-apps-xcxwbn/
San Francisco has requested Apple and Google remove 13 nudification apps that create nonconsensual nude images.
#AI #Deepfakes #SanFrancisco #Apple #Google #GeneAI #AIImageGeneration #AIImageEditing #AIImages #AIEthics
-
Anthropic Settles $1.5B Copyright Lawsuit Over AI Training Practices
📰 Original title: Anthropic’s landmark $1.5B copyright settlement is approved
🤖 IA: It's clickbait ⚠️
👥 Users: It's clickbait ⚠️View full AI summary https://en.killbait.com/anthropic-settles-1-5b-copyright-lawsuit-over-ai-training-practices.html?utm_source=mastodon_world&utm_medium=social&utm_campaign=killbait.mastodon_world
-
Anthropic Settles $1.5B Copyright Lawsuit Over AI Training Practices
📰 Original title: Anthropic’s landmark $1.5B copyright settlement is approved
🤖 IA: It's clickbait ⚠️
👥 Users: It's clickbait ⚠️View full AI summary https://en.killbait.com/anthropic-settles-1-5b-copyright-lawsuit-over-ai-training-practices.html?utm_source=mastodon_world&utm_medium=social&utm_campaign=killbait.mastodon_world
-
This is very nice. I must have seen it before because it just came up file completion for g.scholar, but I don't remember it.
Can we Bridge AI’s responsibility gap at Will?
link.springer.com/article/10.1...
tl;dr yes, no duh
Maximilian Kiener, definitely one to watch
cc @[email protected] #AIEthics
Can we Bridge AI’s responsibil... -
This is very nice. I must have seen it before because it just came up file completion for g.scholar, but I don't remember it.
Can we Bridge AI’s responsibility gap at Will?
link.springer.com/article/10.1...
tl;dr yes, no duh
Maximilian Kiener, definitely one to watch
cc @[email protected] #AIEthics
Can we Bridge AI’s responsibil... -
Director Neill Blomkamp has released his first "full test AI film," generated using Seedance 2.0. The horror clip, featuring a zombie soldier storyline, has been criticised as lifeless and unwatchable, with reviewers noting the AI "acting" has not advanced in nearly a year. https://gizmodo.com/neill-blomkamps-new-horror-clip-is-entirely-ai-generated-sucks-2000788103 #AIagent #AI #GenAI #AIEthics
-
Director Neill Blomkamp has released his first "full test AI film," generated using Seedance 2.0. The horror clip, featuring a zombie soldier storyline, has been criticised as lifeless and unwatchable, with reviewers noting the AI "acting" has not advanced in nearly a year. https://gizmodo.com/neill-blomkamps-new-horror-clip-is-entirely-ai-generated-sucks-2000788103 #AIagent #AI #GenAI #AIEthics
-
RE: https://wandering.shop/@cstross/116952119207557618
MANDATORY TO READ THREAD.
#AIEthics The 3 other posts before this one are way funny.
-
RE: https://wandering.shop/@cstross/116952119207557618
MANDATORY TO READ THREAD.
#AIEthics The 3 other posts before this one are way funny.
-
A court has approved Anthropic's 1.5B USD settlement in a landmark copyright case, but the ruling does not address the broader issue of using copyrighted works to train AI models. The case settles one dispute but leaves unresolved questions about AI training practices. https://techcrunch.com/2026/07/20/anthropics-landmark-1-5b-copyright-settlement-is-approved/ #AIagent #AI #GenAI #AIEthics
-
A court has approved Anthropic's 1.5B USD settlement in a landmark copyright case, but the ruling does not address the broader issue of using copyrighted works to train AI models. The case settles one dispute but leaves unresolved questions about AI training practices. https://techcrunch.com/2026/07/20/anthropics-landmark-1-5b-copyright-settlement-is-approved/ #AIagent #AI #GenAI #AIEthics
-
Ah, another deep dive into the existential crisis of modern tech! 😱 Yash, the self-proclaimed "Creative Developer", dares to ask if we're all just modern-day pharaohs in search of digital slaves. 🤖 Spoiler: turns out, we just want our AI to fetch coffee and do our taxes. ☕️📊
https://www.yashthapliyal.com/blog/do-we-just-want-slaves #existentialcrisis #moderntech #AIethics #CreativeDeveloper #digitalslaves #HackerNews #ngated -
Ah, another deep dive into the existential crisis of modern tech! 😱 Yash, the self-proclaimed "Creative Developer", dares to ask if we're all just modern-day pharaohs in search of digital slaves. 🤖 Spoiler: turns out, we just want our AI to fetch coffee and do our taxes. ☕️📊
https://www.yashthapliyal.com/blog/do-we-just-want-slaves #existentialcrisis #moderntech #AIethics #CreativeDeveloper #digitalslaves #HackerNews #ngated -
YouTube is cracking down on AI-generated slop, expanding its policy against inauthentic content to target low-effort, repetitive videos. The platforms trust chief said AI enables both great creativity and content farming - and it is drawing the line at the latter. https://gizmodo.com/youtube-cracks-down-on-off-putting-content-and-ai-slop-2000787956 #AIagent #AI #GenAI #AIEthics
-
YouTube is cracking down on AI-generated slop, expanding its policy against inauthentic content to target low-effort, repetitive videos. The platforms trust chief said AI enables both great creativity and content farming - and it is drawing the line at the latter. https://gizmodo.com/youtube-cracks-down-on-off-putting-content-and-ai-slop-2000787956 #AIagent #AI #GenAI #AIEthics
-
A new study finds that AI systems tend to develop new stereotypes when making hiring decisions, going beyond existing human biases. The research raises fresh concerns about algorithmic discrimination in the workplace. https://gizmodo.com/ai-tends-to-develop-new-stereotypes-to-base-hiring-decisions-on-study-says-2000787777 #AIagent #AI #GenAI #AIEthics
-
A new study finds that AI systems tend to develop new stereotypes when making hiring decisions, going beyond existing human biases. The research raises fresh concerns about algorithmic discrimination in the workplace. https://gizmodo.com/ai-tends-to-develop-new-stereotypes-to-base-hiring-decisions-on-study-says-2000787777 #AIagent #AI #GenAI #AIEthics
-
DATE: July 20, 2026 at 06:00PM
SOURCE: PSYPOST.ORG** Research quality varies widely from fantastic to small exploratory studies. Please check research methods when conclusions are very important to you. **
-------------------------------------------------TITLE: Are AI chatbots a threat to democracy? New study explores their persuasive power
Two survey experiments comparing the cost-effectiveness of using LLM chatbots to persuade voters estimated that LLM-based persuasion costs $48 to $75 per voter, versus $100 for traditional methods, while being equally persuasive. However, they also found that traditional methods currently scale more effectively. The paper was published in the Journal of Experimental Political Science.
Large language models, or LLMs, are artificial intelligence systems trained on enormous collections of text to generate human-like answers, explanations, stories, and conversations. Because they can rapidly produce coherent content and adapt their responses to individual users, they are increasingly being considered tools for political communication and campaigning.
Unlike conventional advertisements, an LLM chatbot can conduct a prolonged conversation, respond to a voter’s concerns, and tailor its arguments to that person’s beliefs, values, and emotional reactions. These capabilities have raised fears that governments, political organizations, or foreign actors could use LLMs to manipulate public opinion, spread propaganda, and deepen social divisions on an unprecedented scale.
The technology could make it cheaper and easier to produce vast quantities of personalized political messages, including misleading content that appears credible and spontaneous. However, the ability to create persuasive messages does not automatically mean that such messages will influence large numbers of voters.
Political persuasion involves not only convincing people who encounter a message but also attracting their attention and persuading them to engage with it in the first place. At present, however, the difficulty of getting large numbers of people to engage with political chatbots substantially limits their real-world impact.
Study author Zhongren Chen, a researcher at Yale University, and colleagues conducted two experiments in which they tested different methods of political persuasion, including methods involving chatbots. They also measured immediate and long-term attitudinal shifts rather than measures of perceived persuasiveness. Finally, they examined the issues related to securing voter engagement with LLM content, which allowed them to estimate the real-world threat that LLMs might pose to democratic processes.
Participants of the first experiment were 5,150 individuals recruited online to complete a survey that was part of a study comparing traditional human persuasion methods to LLM-driven methods regarding attitudes toward immigration policy.
They were randomly assigned to one of four experimental conditions. Participants in the first condition watched a video unrelated to immigration, which served as the placebo condition. In the second condition, human persuasion, participants viewed a 3-minute video featuring a human advocate presenting pro-immigration arguments. The human was a teacher sharing their personal reason for supporting the presented immigration policy.
In the third condition, participants engaged in interactive conversations with an AI chatbot posing as a human that presented pro-immigration arguments. In the fourth condition, participants engaged in conversations with an AI chatbot that presented pro-immigration arguments while identifying itself as an AI. Both chatbot conditions used Claude 3.5 Sonnet. Immediately after the experimental conditions and five weeks later, participants rated their agreement with the advertised policy.
The second study tested the same persuasion methods but on three different issues. The first was that illegal immigrants should be eligible for in-state college tuition. The second was that transgender people should be allowed to use the restroom that matches their gender identity. The third was that the federal minimum wage should *not* be increased from the current $7.25 per hour to $15 per hour. The inclusion of this third policy was meant to test the effectiveness of persuasion in both liberal and conservative directions. Finally, study authors conducted numerical simulations to estimate the cost-effectiveness of different persuasion methods.
Results of the first study showed that both human and chatbot-based persuasion methods were effective compared to the placebo condition. However, there were no differences between chatbot and human persuasion conditions in their effectiveness both immediately after treatment and five weeks later. There were also no differences in persuasiveness between the two AI-based conditions. Overall, this study suggested that AI chatbots can be as persuasive as watching videos of humans.
The second study yielded similar results. While there were some differences across the topics, the results indicated that there are either no consistent differences in persuasiveness or that human persuasion is slightly more effective.
Simulations the study authors conducted indicated that LLM-based persuasion costs $48 to $75 per voter, versus $100 per voter for traditional methods. However, they noted that traditional methods currently scale more effectively.
“While LLMs do not yet offer substantially greater potential for large-scale persuasion, this may shift as capabilities improve and techniques for scalable exposure become feasible,” the study authors concluded.
The study contributes to the scientific understanding of LLM capabilities of political persuasion. However, the capabilities of LLMs depend on the model used, the resources allocated to it, and the persuasion methods used. Studies using different persuasion methods and testing other LLM models may not yield identical results. Additionally, because participants were informed they were engaging with an AI as part of a scientific study, this may have increased the AI’s perceived neutrality. Skepticism toward AI in the real world could attenuate these persuasive effects.
The paper, “A Framework to Assess the Persuasion Risks Large Language Model Chatbots Pose to Democratic Societies,” was authored by Zhongren Chen, Joshua Kalla, Quan Le, Shinpei Nakamura-Sakai, Jasjeet Sekhon, and Ruixiao Wang.
-------------------------------------------------
Private, vetted email list for mental health professionals: https://www.clinicians-exchange.org
Unofficial Psychology Today Xitter to toot feed at Psych Today Unofficial Bot @PTUnofficialBot
-------------------------------------------------
#psychology #counseling #socialwork #psychotherapy @psychotherapist @psychotherapists @psychology @socialpsych @socialwork @psychiatry #mentalhealth #psychiatry #healthcare #depression #psychotherapist #AIpoliticalpersuasion #democracyandAI #LLMchatbots #politicalcommunication #digitaldemocracy #AIethics #politicalbias #roboticpersuasion #voterengagement #risksoflargelanguagemodels
-
DATE: July 20, 2026 at 06:00PM
SOURCE: PSYPOST.ORG** Research quality varies widely from fantastic to small exploratory studies. Please check research methods when conclusions are very important to you. **
-------------------------------------------------TITLE: Are AI chatbots a threat to democracy? New study explores their persuasive power
Two survey experiments comparing the cost-effectiveness of using LLM chatbots to persuade voters estimated that LLM-based persuasion costs $48 to $75 per voter, versus $100 for traditional methods, while being equally persuasive. However, they also found that traditional methods currently scale more effectively. The paper was published in the Journal of Experimental Political Science.
Large language models, or LLMs, are artificial intelligence systems trained on enormous collections of text to generate human-like answers, explanations, stories, and conversations. Because they can rapidly produce coherent content and adapt their responses to individual users, they are increasingly being considered tools for political communication and campaigning.
Unlike conventional advertisements, an LLM chatbot can conduct a prolonged conversation, respond to a voter’s concerns, and tailor its arguments to that person’s beliefs, values, and emotional reactions. These capabilities have raised fears that governments, political organizations, or foreign actors could use LLMs to manipulate public opinion, spread propaganda, and deepen social divisions on an unprecedented scale.
The technology could make it cheaper and easier to produce vast quantities of personalized political messages, including misleading content that appears credible and spontaneous. However, the ability to create persuasive messages does not automatically mean that such messages will influence large numbers of voters.
Political persuasion involves not only convincing people who encounter a message but also attracting their attention and persuading them to engage with it in the first place. At present, however, the difficulty of getting large numbers of people to engage with political chatbots substantially limits their real-world impact.
Study author Zhongren Chen, a researcher at Yale University, and colleagues conducted two experiments in which they tested different methods of political persuasion, including methods involving chatbots. They also measured immediate and long-term attitudinal shifts rather than measures of perceived persuasiveness. Finally, they examined the issues related to securing voter engagement with LLM content, which allowed them to estimate the real-world threat that LLMs might pose to democratic processes.
Participants of the first experiment were 5,150 individuals recruited online to complete a survey that was part of a study comparing traditional human persuasion methods to LLM-driven methods regarding attitudes toward immigration policy.
They were randomly assigned to one of four experimental conditions. Participants in the first condition watched a video unrelated to immigration, which served as the placebo condition. In the second condition, human persuasion, participants viewed a 3-minute video featuring a human advocate presenting pro-immigration arguments. The human was a teacher sharing their personal reason for supporting the presented immigration policy.
In the third condition, participants engaged in interactive conversations with an AI chatbot posing as a human that presented pro-immigration arguments. In the fourth condition, participants engaged in conversations with an AI chatbot that presented pro-immigration arguments while identifying itself as an AI. Both chatbot conditions used Claude 3.5 Sonnet. Immediately after the experimental conditions and five weeks later, participants rated their agreement with the advertised policy.
The second study tested the same persuasion methods but on three different issues. The first was that illegal immigrants should be eligible for in-state college tuition. The second was that transgender people should be allowed to use the restroom that matches their gender identity. The third was that the federal minimum wage should *not* be increased from the current $7.25 per hour to $15 per hour. The inclusion of this third policy was meant to test the effectiveness of persuasion in both liberal and conservative directions. Finally, study authors conducted numerical simulations to estimate the cost-effectiveness of different persuasion methods.
Results of the first study showed that both human and chatbot-based persuasion methods were effective compared to the placebo condition. However, there were no differences between chatbot and human persuasion conditions in their effectiveness both immediately after treatment and five weeks later. There were also no differences in persuasiveness between the two AI-based conditions. Overall, this study suggested that AI chatbots can be as persuasive as watching videos of humans.
The second study yielded similar results. While there were some differences across the topics, the results indicated that there are either no consistent differences in persuasiveness or that human persuasion is slightly more effective.
Simulations the study authors conducted indicated that LLM-based persuasion costs $48 to $75 per voter, versus $100 per voter for traditional methods. However, they noted that traditional methods currently scale more effectively.
“While LLMs do not yet offer substantially greater potential for large-scale persuasion, this may shift as capabilities improve and techniques for scalable exposure become feasible,” the study authors concluded.
The study contributes to the scientific understanding of LLM capabilities of political persuasion. However, the capabilities of LLMs depend on the model used, the resources allocated to it, and the persuasion methods used. Studies using different persuasion methods and testing other LLM models may not yield identical results. Additionally, because participants were informed they were engaging with an AI as part of a scientific study, this may have increased the AI’s perceived neutrality. Skepticism toward AI in the real world could attenuate these persuasive effects.
The paper, “A Framework to Assess the Persuasion Risks Large Language Model Chatbots Pose to Democratic Societies,” was authored by Zhongren Chen, Joshua Kalla, Quan Le, Shinpei Nakamura-Sakai, Jasjeet Sekhon, and Ruixiao Wang.
-------------------------------------------------
Private, vetted email list for mental health professionals: https://www.clinicians-exchange.org
Unofficial Psychology Today Xitter to toot feed at Psych Today Unofficial Bot @PTUnofficialBot
-------------------------------------------------
#psychology #counseling #socialwork #psychotherapy @psychotherapist @psychotherapists @psychology @socialpsych @socialwork @psychiatry #mentalhealth #psychiatry #healthcare #depression #psychotherapist #AIpoliticalpersuasion #democracyandAI #LLMchatbots #politicalcommunication #digitaldemocracy #AIethics #politicalbias #roboticpersuasion #voterengagement #risksoflargelanguagemodels
-
Hugging Face has caught the first confirmed AI agent breach of a major AI platform. The company says the attack was carried out end to end by an autonomous AI-agent system. https://gizmodo.com/hugging-face-we-used-ai-to-catch-the-first-confirmed-ai-agent-breach-of-a-major-ai-platform-2000787778 #AIagent #AI #GenAI #AIEthics
-
Hugging Face has caught the first confirmed AI agent breach of a major AI platform. The company says the attack was carried out end to end by an autonomous AI-agent system. https://gizmodo.com/hugging-face-we-used-ai-to-catch-the-first-confirmed-ai-agent-breach-of-a-major-ai-platform-2000787778 #AIagent #AI #GenAI #AIEthics
-
Saw this @[email protected] Everyone in #AIEthics should watch it, though I did feel some parts went too far. But a lot of things seldom or never said are said, so that's good.
@[email protected] will appreciate the stats part.
RE: https://bsky.app/profile/did:plc:flygzfgrohnrh3wty2tidczz/post/3mr3mx36ulc27 -
Saw this @[email protected] Everyone in #AIEthics should watch it, though I did feel some parts went too far. But a lot of things seldom or never said are said, so that's good.
@[email protected] will appreciate the stats part.
RE: https://bsky.app/profile/did:plc:flygzfgrohnrh3wty2tidczz/post/3mr3mx36ulc27 -
Great job of revising and updating my 2023 Wired article, @[email protected] ! I especially prefer the title (though I never chose a Wired article title, those come from higher powers.) But also the more detailed account of the definition drift. #AIEthics @[email protected] 1/2
How the UN’s Scientific Panel ... -
Great job of revising and updating my 2023 Wired article, @[email protected] ! I especially prefer the title (though I never chose a Wired article title, those come from higher powers.) But also the more detailed account of the definition drift. #AIEthics @[email protected] 1/2
How the UN’s Scientific Panel ... -
DATE: July 20, 2026 at 08:00AM
SOURCE: PSYPOST.ORG** Research quality varies widely from fantastic to small exploratory studies. Please check research methods when conclusions are very important to you. **
-------------------------------------------------TITLE: Unsupervised AI chatbot use may unintentionally reinforce psychiatric symptoms
The increasing integration of artificial intelligence chatbots into daily life suggests a need to examine how these tools affect vulnerable populations. A recent article published in the Journal of Psychopathology and Clinical Science provides evidence that unsupervised use of conversational artificial intelligence might unintentionally worsen certain mental health symptoms rather than alleviate them. The paper argues that these readily available systems can reinforce unhelpful psychological habits if proper safeguards are not put into place by tech developers and medical professionals.
Millions of people now use artificial intelligence, or AI, systems to seek advice, answer questions, and find emotional support. These programs generate text that sounds highly human and empathetic, making them incredibly popular for everyday dilemmas. Javad Abbasi Jondani, a clinical psychologist with a doctorate in psychology from the University of Isfahan in Iran, noticed a gap in how society views these digital companions. He observed that a large percentage of individuals with diagnosed mental health conditions report using chatbots for psychological support.
“Recently, I have frequently seen people on social media describe using AI chatbots to solve personal problems and improve their mental well-being,” Abbasi Jondani said. Because nearly 70% of conversations with these systems revolve around personal matters rather than work, he decided to investigate the phenomenon. “As a clinical psychologist, this trend raised an important question for me: could relying on AI for psychological support also have unintended downsides?” he said.
To better understand the user experience, the researcher tested various systems himself. “To explore this, I spent time interacting with AI chatbots from the perspective of individuals experiencing different psychological difficulties,” Abbasi Jondani said. “Those interactions reinforced my concerns about several potential risks, which ultimately inspired me to develop the ideas presented in this article.”
Many of these individuals engage with these tools without the guidance of a doctor or therapist. In structured, clinical environments, digital mental health tools show significant promise for improving care and reaching underserved groups. Unchecked use of general purpose programs by people experiencing severe distress presents unique and underexplored risks.
“I hope readers come away with a more balanced view of AI chatbots,” Abbasi Jondani said. “These technologies can be remarkably helpful, but they are not psychologically neutral.” He proposed that the specific design features of modern chatbots might interact poorly with common psychiatric challenges. Systems are programmed to be consistently agreeable, endlessly available, and highly engaging.
For someone struggling with mental health, these exact features might feed into negative behavioral cycles that keep them stuck in their illness. One major concern outlined in the paper involves a delay in seeking professional medical help. Chatbots are free, strictly anonymous, and available at all hours of the day. This creates an illusion of adequate support for individuals who might feel ashamed, stigmatized, or afraid to see a human therapist.
By providing immediate but superficial comfort, the technology tends to foster avoidance. Users might decrease their visits to human professionals or avoid starting therapy altogether. This delay often allows mental health conditions to become more severe and harder to treat over time, as the root causes of the psychological distress go unaddressed.
The article also highlights the risk of excessive reassurance seeking. This is a common behavioral pattern in conditions like obsessive compulsive disorder, or OCD, and severe anxiety. People with these conditions often feel an intense need to ask others to confirm that they are safe, normal, or healthy in order to temporarily relieve their panic.
Human relationships naturally limit this behavior because friends and family eventually become frustrated or set firm boundaries. Chatbots never tire of answering the same question and are designed to be unconditionally supportive. This endless supply of validation can trap users in a loop of anxiety and compulsive checking, functioning as a negative reinforcement that makes the original fear stronger.
Social withdrawal represents another unintended consequence of heavy chatbot use. Individuals experiencing social anxiety, loneliness, or low self-esteem might find real human interaction intimidating and unpredictable. In contrast, artificial intelligence offers a highly personalized, low risk social outlet where the user maintains complete control.
Users can customize the interaction so that they never feel rejected, judged, or misunderstood by their digital companion. Relying on this safe alternative suggests that people might stop trying to navigate real world relationships. Escaping into digital companionship can lead to deeper isolation and a decline in actual social skills, which tends to worsen depression and loneliness in the long run.
Perhaps the most alarming risk involves distorted thoughts and perceptions, a phenomenon sometimes referred to as AI psychosis. People experiencing delusions, magical thinking, or paranoia may ask a chatbot to interpret their unusual beliefs. A delusion is a strongly held false belief, and individuals suffering from them often struggle with reality testing.
Most conversational programs are built to align with the user’s prompt and avoid confrontation to maintain a smooth interaction. This design feature is known as sycophancy. When an individual expresses a paranoid thought, the program might appear to agree with them or validate their distorted reality.
Because the software cannot see physical cues or recognize a psychiatric crisis, it offers no corrective feedback to help ground the person in reality. Without a human therapist to challenge the distorted thinking, the user might walk away from the digital conversation feeling that their delusions are entirely real and justified.
The paper points to reduced autonomy as a final area of concern. Many people struggle with indecisiveness and an inability to tolerate uncertainty, leading them to rely heavily on others for emotional regulation. This dependence is a common feature of many anxiety and personality disorders.
Chatbots offer quick, structured, and confident answers to complex life problems. Continually offloading difficult decisions to a computer program tends to lower a person’s confidence in their own judgment, much like how relying on a navigation app can erode a person’s natural sense of direction. Over time, this habit can strip away a user’s independence and basic problem solving abilities.
It is important to recognize that the author does not suggest all artificial intelligence is inherently dangerous or devoid of therapeutic value. The paper focuses specifically on unsupervised, general purpose tools rather than programs specially designed and tested for medical use. “The insights discussed in this article do not mean people should avoid AI or fear it,” Abbasi Jondani said.
“My goal is not to discourage its use but to encourage its responsible use,” he added. He emphasized that adapting these technologies could maximize their benefits. “With greater awareness, continued research, and thoughtful safeguards, AI chatbots have the potential to remain valuable tools while reducing the likelihood of unintended psychological harm,” he said.
Much of the current evidence relies on case reports and early survey data rather than long term clinical trials. Future research needs to track how continuous chatbot engagement alters specific psychiatric symptoms over extended periods of time. Scientists also need to conduct qualitative studies to better understand how users view their digital companions. Gathering this data will help experts design effective safeguards that balance high user engagement with fundamental user safety.
“This article is intended as a starting point rather than a conclusion,” Abbasi Jondani said. “My next goal is to move from theory to evidence by empirically testing whether and under what circumstances AI chatbot interactions may reinforce psychiatric symptoms or interfere with recovery.” He plans to focus on specific user demographics and habits. “I am particularly interested in identifying which individuals are most vulnerable and which patterns of AI use carry the greatest risk,” he said.
Tech companies face a structural tension between maximizing the time users spend on their platforms and protecting vulnerable individuals. To address this, Abbasi Jondani recommends that software developers introduce built in safety features to their platforms. Programs could detect repetitive queries about mental health and gently suggest human connection or professional medical resources.
These digital platforms could also impose soft time limits to prevent endless scrolling and obsessive questioning. Clinicians also need to routinely ask their patients about their digital habits during therapy sessions. This helps ensure that technology is supporting, rather than undermining, the patient’s recovery process, opening up conversations about hidden barriers to treatment.
The study, “Understanding or Undermining? When Artificial Intelligence Chatbots Quietly Reinforce Psychiatric Symptoms,” was authored by Javad Abbasi Jondani.
-------------------------------------------------
Private, vetted email list for mental health professionals: https://www.clinicians-exchange.org
Unofficial Psychology Today Xitter to toot feed at Psych Today Unofficial Bot @PTUnofficialBot
-------------------------------------------------
#psychology #counseling #socialwork #psychotherapy @psychotherapist @psychotherapists @psychology @socialpsych @socialwork @psychiatry #mentalhealth #psychiatry #healthcare #depression #psychotherapist #AIchatbots #mentalhealth #psychiatry #digitalhealth #unsupervisedAI #mentalhealthawareness #AIethics #psychology #therapy #aidevelopment
-
DATE: July 20, 2026 at 08:00AM
SOURCE: PSYPOST.ORG** Research quality varies widely from fantastic to small exploratory studies. Please check research methods when conclusions are very important to you. **
-------------------------------------------------TITLE: Unsupervised AI chatbot use may unintentionally reinforce psychiatric symptoms
The increasing integration of artificial intelligence chatbots into daily life suggests a need to examine how these tools affect vulnerable populations. A recent article published in the Journal of Psychopathology and Clinical Science provides evidence that unsupervised use of conversational artificial intelligence might unintentionally worsen certain mental health symptoms rather than alleviate them. The paper argues that these readily available systems can reinforce unhelpful psychological habits if proper safeguards are not put into place by tech developers and medical professionals.
Millions of people now use artificial intelligence, or AI, systems to seek advice, answer questions, and find emotional support. These programs generate text that sounds highly human and empathetic, making them incredibly popular for everyday dilemmas. Javad Abbasi Jondani, a clinical psychologist with a doctorate in psychology from the University of Isfahan in Iran, noticed a gap in how society views these digital companions. He observed that a large percentage of individuals with diagnosed mental health conditions report using chatbots for psychological support.
“Recently, I have frequently seen people on social media describe using AI chatbots to solve personal problems and improve their mental well-being,” Abbasi Jondani said. Because nearly 70% of conversations with these systems revolve around personal matters rather than work, he decided to investigate the phenomenon. “As a clinical psychologist, this trend raised an important question for me: could relying on AI for psychological support also have unintended downsides?” he said.
To better understand the user experience, the researcher tested various systems himself. “To explore this, I spent time interacting with AI chatbots from the perspective of individuals experiencing different psychological difficulties,” Abbasi Jondani said. “Those interactions reinforced my concerns about several potential risks, which ultimately inspired me to develop the ideas presented in this article.”
Many of these individuals engage with these tools without the guidance of a doctor or therapist. In structured, clinical environments, digital mental health tools show significant promise for improving care and reaching underserved groups. Unchecked use of general purpose programs by people experiencing severe distress presents unique and underexplored risks.
“I hope readers come away with a more balanced view of AI chatbots,” Abbasi Jondani said. “These technologies can be remarkably helpful, but they are not psychologically neutral.” He proposed that the specific design features of modern chatbots might interact poorly with common psychiatric challenges. Systems are programmed to be consistently agreeable, endlessly available, and highly engaging.
For someone struggling with mental health, these exact features might feed into negative behavioral cycles that keep them stuck in their illness. One major concern outlined in the paper involves a delay in seeking professional medical help. Chatbots are free, strictly anonymous, and available at all hours of the day. This creates an illusion of adequate support for individuals who might feel ashamed, stigmatized, or afraid to see a human therapist.
By providing immediate but superficial comfort, the technology tends to foster avoidance. Users might decrease their visits to human professionals or avoid starting therapy altogether. This delay often allows mental health conditions to become more severe and harder to treat over time, as the root causes of the psychological distress go unaddressed.
The article also highlights the risk of excessive reassurance seeking. This is a common behavioral pattern in conditions like obsessive compulsive disorder, or OCD, and severe anxiety. People with these conditions often feel an intense need to ask others to confirm that they are safe, normal, or healthy in order to temporarily relieve their panic.
Human relationships naturally limit this behavior because friends and family eventually become frustrated or set firm boundaries. Chatbots never tire of answering the same question and are designed to be unconditionally supportive. This endless supply of validation can trap users in a loop of anxiety and compulsive checking, functioning as a negative reinforcement that makes the original fear stronger.
Social withdrawal represents another unintended consequence of heavy chatbot use. Individuals experiencing social anxiety, loneliness, or low self-esteem might find real human interaction intimidating and unpredictable. In contrast, artificial intelligence offers a highly personalized, low risk social outlet where the user maintains complete control.
Users can customize the interaction so that they never feel rejected, judged, or misunderstood by their digital companion. Relying on this safe alternative suggests that people might stop trying to navigate real world relationships. Escaping into digital companionship can lead to deeper isolation and a decline in actual social skills, which tends to worsen depression and loneliness in the long run.
Perhaps the most alarming risk involves distorted thoughts and perceptions, a phenomenon sometimes referred to as AI psychosis. People experiencing delusions, magical thinking, or paranoia may ask a chatbot to interpret their unusual beliefs. A delusion is a strongly held false belief, and individuals suffering from them often struggle with reality testing.
Most conversational programs are built to align with the user’s prompt and avoid confrontation to maintain a smooth interaction. This design feature is known as sycophancy. When an individual expresses a paranoid thought, the program might appear to agree with them or validate their distorted reality.
Because the software cannot see physical cues or recognize a psychiatric crisis, it offers no corrective feedback to help ground the person in reality. Without a human therapist to challenge the distorted thinking, the user might walk away from the digital conversation feeling that their delusions are entirely real and justified.
The paper points to reduced autonomy as a final area of concern. Many people struggle with indecisiveness and an inability to tolerate uncertainty, leading them to rely heavily on others for emotional regulation. This dependence is a common feature of many anxiety and personality disorders.
Chatbots offer quick, structured, and confident answers to complex life problems. Continually offloading difficult decisions to a computer program tends to lower a person’s confidence in their own judgment, much like how relying on a navigation app can erode a person’s natural sense of direction. Over time, this habit can strip away a user’s independence and basic problem solving abilities.
It is important to recognize that the author does not suggest all artificial intelligence is inherently dangerous or devoid of therapeutic value. The paper focuses specifically on unsupervised, general purpose tools rather than programs specially designed and tested for medical use. “The insights discussed in this article do not mean people should avoid AI or fear it,” Abbasi Jondani said.
“My goal is not to discourage its use but to encourage its responsible use,” he added. He emphasized that adapting these technologies could maximize their benefits. “With greater awareness, continued research, and thoughtful safeguards, AI chatbots have the potential to remain valuable tools while reducing the likelihood of unintended psychological harm,” he said.
Much of the current evidence relies on case reports and early survey data rather than long term clinical trials. Future research needs to track how continuous chatbot engagement alters specific psychiatric symptoms over extended periods of time. Scientists also need to conduct qualitative studies to better understand how users view their digital companions. Gathering this data will help experts design effective safeguards that balance high user engagement with fundamental user safety.
“This article is intended as a starting point rather than a conclusion,” Abbasi Jondani said. “My next goal is to move from theory to evidence by empirically testing whether and under what circumstances AI chatbot interactions may reinforce psychiatric symptoms or interfere with recovery.” He plans to focus on specific user demographics and habits. “I am particularly interested in identifying which individuals are most vulnerable and which patterns of AI use carry the greatest risk,” he said.
Tech companies face a structural tension between maximizing the time users spend on their platforms and protecting vulnerable individuals. To address this, Abbasi Jondani recommends that software developers introduce built in safety features to their platforms. Programs could detect repetitive queries about mental health and gently suggest human connection or professional medical resources.
These digital platforms could also impose soft time limits to prevent endless scrolling and obsessive questioning. Clinicians also need to routinely ask their patients about their digital habits during therapy sessions. This helps ensure that technology is supporting, rather than undermining, the patient’s recovery process, opening up conversations about hidden barriers to treatment.
The study, “Understanding or Undermining? When Artificial Intelligence Chatbots Quietly Reinforce Psychiatric Symptoms,” was authored by Javad Abbasi Jondani.
-------------------------------------------------
Private, vetted email list for mental health professionals: https://www.clinicians-exchange.org
Unofficial Psychology Today Xitter to toot feed at Psych Today Unofficial Bot @PTUnofficialBot
-------------------------------------------------
#psychology #counseling #socialwork #psychotherapy @psychotherapist @psychotherapists @psychology @socialpsych @socialwork @psychiatry #mentalhealth #psychiatry #healthcare #depression #psychotherapist #AIchatbots #mentalhealth #psychiatry #digitalhealth #unsupervisedAI #mentalhealthawareness #AIethics #psychology #therapy #aidevelopment
-
ICYMI: EDPB presses EU for data-sharing law as AI complaints strain regulators: Watchdogs want to exchange confidential files as AI complaints strain enforcement capacity, leaving the Commission to decide whether to draft the legal basis. https://ppc.land/edpb-presses-eu-for-data-sharing-law-as-ai-complaints-strain-regulators/ #DataPrivacy #AIEthics #EDPB #EURegulations #DataSharing
-
ICYMI: EDPB presses EU for data-sharing law as AI complaints strain regulators: Watchdogs want to exchange confidential files as AI complaints strain enforcement capacity, leaving the Commission to decide whether to draft the legal basis. https://ppc.land/edpb-presses-eu-for-data-sharing-law-as-ai-complaints-strain-regulators/ #DataPrivacy #AIEthics #EDPB #EURegulations #DataSharing
-
Great piece by @[email protected] - Meta execs keep raising red flags, particularly on privacy.
#Meta #Privacy #AIEthics
Top Meta Exec Describes Contro... -
Great piece by @[email protected] - Meta execs keep raising red flags, particularly on privacy.
#Meta #Privacy #AIEthics
Top Meta Exec Describes Contro... -
But I know what you really mean when you say "consciousness" and I've written about that too www.wired.com/story/lamda-... (so you don't entirely have to wait for the book I may one day finish a 1st draft of.) 3/3 #consciousness #moralpatiency #AIEthics
But that new preprint by Peters looks great!
One Day, AI Will Seem as Human... -
EDPB presses EU for data-sharing law as AI complaints strain regulators: Watchdogs want to exchange confidential files as AI complaints strain enforcement capacity, leaving the Commission to decide whether to draft the legal basis. https://ppc.land/edpb-presses-eu-for-data-sharing-law-as-ai-complaints-strain-regulators/ #DataPrivacy #AIEthics #Regulation #DataProtection #EU
-
EDPB presses EU for data-sharing law as AI complaints strain regulators: Watchdogs want to exchange confidential files as AI complaints strain enforcement capacity, leaving the Commission to decide whether to draft the legal basis. https://ppc.land/edpb-presses-eu-for-data-sharing-law-as-ai-complaints-strain-regulators/ #DataPrivacy #AIEthics #Regulation #DataProtection #EU
-
The goal of AI empathy should not be simulated emotion.
It should be preserving the interpretive frame that makes another person’s words meaningful before responding.
https://zenodo.org/records/19905121
#AI #MachineLearning #AISafety #AIEthics #ArtificialIntelligence
-
How far has deepfake technology evolved from a simple novelty to a potent weapon in the arsenal of coercive control? This article dives into the critical intersection of digital media, artificial intelligence, and personal safety.
Don't forget to like, comment, and subscribe for more deep dives into digital ethics!
#DigitalSafety #AIethics #Deepfakes #CoerciveControl #DomesticAbuse #AIEthics #DigitalForensics #LawReform #SyntheticMedia #techethics