One of these studies focuses on the AI Chatbot Behavior Change Model to guide patients through behavioral changes and promote healthy choices. The affordances of technology such as chatbots is that they may be scalable to large populations. People who experience difficulty accessing resources may be able to access AI chatbots electronically. AI may also provide people a form of social support and a sense of self-efficacy. Deep learning programs are continuously adapting to the environment and the person, so a great deal of personalization is possible. The chatbots are also available at all times unlike physicians or other paid workers (Oh et. al, 2021). This constant accessibility may allow behavioral change programs to be more effective because the patients can seek help in times of difficulty or when their motivation is lacking. This deep personalization can prevent the problems associated with rigid medical protocols and can direct the individual in the direction optimal for his or her growth. The possibility of scalability may allow for health change on a larger level than ever possible before.     

Many AI chatbots have the ability to disseminate health related information and potentially influence the behaviors of humans. Different Chatbots vary in their abilities, characteristics, and affordances to the people using them. A systematic review was conducted to compare these functions, and determine the effects of these chatbots on health behaviors such as exercise, healthy diets, and weight maintenance. AI chatbots may also affect long term health outcomes. Five out of the seven studies found that the chatbots effectively increased the physical exercise of the users. Diet and weight differences were not clear (Oh et. al, 2021). While the research is not clear up to this point, there does seem to be promise in the field of behavioral change. The field of AI chatbot lifestyle and behavior modification is expected to expand rapidly in the near future (Oh et. al, 2021). The technology is rapidly changing, so the current applications and future applications of AI in the field of behavioral change are dynamic and ever improving. Current technologies are offered online to help people maintain medical protocols and behaviors. For example, the program Intercom is a platform design to help patients with medication and making appointments.

Physical health is only one way that AI technology can benefit individuals. According to a review that outlines the field of cyber psychology, AI and robots-human interactions will become an important part of the mental health field. AI has already begun use in clinical settings in the form of robo-therapy. AI technology and robots can be used to assist psychotherapists and can help diagnose patients. AI has already been adapted to help children with autism spectrum disorder. Social fluent robots can help children develop their social skills (Ancis et.al., 2020). While robots of that past have been rigid and limited in their communication complexity, AI has provided new ways for humans and robots to interact. Robots that once seemed cold and lifeless now have impressive abilities to imitate human speech. These abilities allow new systems to function socially. This has led to the engineering of companion based robots.

A seal robot named Paro was built to provide older populations with social support. Paro’s social support promotes mental health by increasing social connection and reduces loneliness. Robots like Paro encourage people and improve their mood. Paro has also been found to reduce long term stress in elderly people. Robots like Paro have also begun use in pediatric facilities as well (Ancis et.al., 2020). These social functions are expanding rapidly with new language learning models of AI. With the well documented effects of social support on health, the potential for AI assisted social support may greatly aid mental health. This is especially true for potentially isolated populations.

     AI artificial intelligence also has the potential to influence mental health through therapy. The mental health field has already introduced AI therapy assistance to help people cope with psychological disorders. One of the most prevalent psychological disorders is anxiety. Anxiety has one of the highest prevalence rates of any psychological disorder and there is a high demand for therapists. This means that additional help is needed to provide psychotherapy for patients. AI assisted psychotherapy provides a possible solution. If successful, AI therapy may provide standardized forms of psychotherapy that can become available to those who need it. Previous research has shown that online micro-video therapy can promote resilience in those with anxiety and depression, but I may also effectively deliver CBT. Computerized CBT has already been used to help those with OCD and substance use disorders. In an ongoing study an AI chatbot, XIAO AN, was designed to deliver CBT therapy. It also has the capability to recognize mood in patients. It was designed to supplement therapists because of the deficit in care (SU, et. al., 2022). If robo-therapy and AI chatbots are able to provide effective therapy as some research indicates, then therapists may be able to focus more on the individualized needs of patients. The complex task of CBT may soon be available to a large population that would not attend traditional therapy.

As AI becomes more human-like, its abilities and roles mimic those of humans. One way that AI has begun to do this is through conversational ability. Not only can AI provide needed health advice and information, but conversational ability allows for AI to be incorporated into therapy, social support and other human focused domains. This leads to the question of how this technology is going to be incorporated into these human domains.

For example, conversational artificial intelligence is increasing in use in healthcare. As shown before, some AI can effectively acquire diagnostic information and it can deliver evidence based psychotherapy. AI can also give clinicians advice and feedback on their therapy. AI chats also have the can talk to young people about sensitive issues such as sex, suicide, and drug use (Miner et.al., 2019). This may be a good thing if standardized but it raises some concerns due to the variation in AI coding and the risk of using a platform without human sensibilities.

At the moment most AI is used collaboratively with a therapist. One perspective is that this is the maximum extent that it should be used. Some people see conversational AI is an unnecessary risk to patients and its use should be avoided in case of unintended consequences. Another perspective on AI proposes AI presents a unique opportunity to help patients. AI could individually personalize itself to match a patient's needs in a way psychotherapists are unable. It may also provide access to people who could not get access to a clinician (Miner et.al., 2019). Balancing these concerns and benefits are important for the integration of AI into the mental health field.

One way that AI could be integrated is by using AI only to inform psychotherapists. This would not increase access, but it may improve the quality of therapy. Another option is that AI delivers the therapy and humans supervise. This would slightly increase access, but would disrupt the client clinician relationship. The most extreme form of integration is AI alone therapy. This therapy is highly scalable because it does not rely on the attention and presence of a human. However, there is no patient-psychotherapist relationship, and the quality and patient self-disclosure is also unknown (Miner et.al., 2019). These three methods vary by risk. While the latter options may provide sweeping benefits, they also present disruptions to the field.

However, conversational AI in the health field remains highly optimistic. Patients with PTSD symptoms who have interacted with conversational AI have responded with high acceptance. AI has been studied in many clinical groups such as clinical depression and student stress, however much more research is still needed. As healthcare continues AI integration it is important that safety, trust, and oversight are evaluated throughout the process. It is important that the healthcare sector and the tech sector are wary of potential problems and work together cooperatively and effectively (Miner et.al., 2019). Reflective studies and AI evaluations are important safeguards against unintended effects of highly influential AI programs in the health field.