Artificial Intelligence in Clinical Practice:

This article will focus on the use of Artificial Intelligence (AI) within an Electronic Health Record (EHR).  Primarily, AI is utilized to assist in writing progress notes when tracking client care.  There are other uses of AI including AI Robots and other tools, but this article will not address those applications.

When considering the use of AI there are several considerations related to privacy and ethical practices.  Related to confidentiality, AI utilizes a microphone to record sessions.  These recording sessions are the basis of creating progress notes about the interaction between the client and the counselor.  The clinician is required to review the AI generated progress note to confirm it’s accuracy.  Policies and procedures must be created related to privacy and recorded information.  These policies and procedures will require a higher standard than simply protecting written information within the EHR. In the article, Updating HIPAA Security to Respond to Artificial Intelligence (November 13, 2023) it stated, “Organizations need to improve data security, reduce the risks of cyber threats, and maintain constant vigilance for potential weaknesses.”  They reiterate the importance of updating policies and procedures and ensuring that all staff is trained related to AI.  The article also discusses the importance of client safety being “paramount and non-maleficence” which is doing no harm to those we serve.

Additionally, updating your Consent to Treatment form is a critical step. It is important to add the risks and benefits of AI to the Consent to Treatment language.  The client should be made aware that AI is being utilized.  Additionally, the client needs to explicitly consent to the utilization of AI including the digital recording of sessions.

The American Counseling Association in the article, Recommendations for Client Use and Caution of Artificial Intelligence recommended that AI not be utilized for the diagnosis of clients. They noted, “while a powerful tool, it may not fully capture the nuanced understanding and clinical judgement required for accurate mental health diagnoses. Unlike human counselors, AI lacks the ability to holistically consider a client’s complex personal history, cultural context, and varied symptoms and factors among others (Kulkarni & Sigh, 2023)

It is also important to confirm with the state Licensing Board who credentials professionals that AI is an allowable tool related to writing progress notes and other possible applications.  The question for many clinicians should be, can I defend myself if there is a complaint to the board related to my clinical practice? Also, would you be able to defend yourself in a court of law if challenged regarding something that was written in a progress note assisted by AI? In the article, Updating HIPAA Security to Respond to Artificial Intelligence, it stated “If it (AI) is used, then who’s liable if AI and the provider disagree.  What harm is incurred?”

The American Counseling Association also cautioned the use of AI related to cultural competency.  “AI systems often rely on data that may not adequately represent all communities, particularly marginalized groups” (Celi et al., 2022).  We see this as well in the Diagnostic Statistical Manual 5TR. When diagnosing, we must take into consideration a person’s cultural background understanding that the DSM 5TR is based on a limited dataset culturally.

Lastly, staff who utilize AI must be trained to understand the limitations of its use and the importance of review and critique of any AI generated documentation.  The clinician cannot just simply approve the note without this review.  AI does have the risk of clinicians taking a short cut to completing the paperwork without adequately reviewing the documentation for accuracy. AI has limitations when compared to the human brain. AI lacks the ability to intuitively pick up on human expressions, verbal and nonverbal responses and communication for example.

A provider can improve the use of AI within an EHR with concrete policies, training and quality assurance practices that assure the staff using these tools are doing it accurately with the client’s best interest in mind.  It is recommended when reviewing potential EHR’s that utilize AI, that you ask questions related to the ethics and privacy of these tools.

Related Blog Posts

Related Learning Labs

Related Resources