Artificial intelligence in mental health care

Artificial intelligence (AI) is used across industries and is integrated into many of our daily activities, such as banking, social media, online shopping, streaming services, and more. AI is an umbrella term that encompasses several types of technology. While there is no single, universally accepted definition, AI is generally defined as a subdiscipline of computer science that aims to produce programs that simulate human intelligence.

AI has many subfields, some of which are shown below. These subfields often overlap and are interrelated.

Read an overview of these key terms.

Increasingly, AI is being utilized in health care settings. AI is used to streamline administrative tasks, make workflows more efficient, and aid in clinical decision-making. And the possible uses for AI are projected to only grow.

Administrative use

Given the ever-increasing demands on clinicians’ time, there is great interest in using AI tools to streamline administrative tasks in the hopes of decreasing burnout and increasing time for patient care. Some of the processes where AI could be employed are:

• Automating scheduling and appointment reminders
• Streamlining routine communication—such as providing educational information or answers to frequently asked questions
• Generating clinical notes
• Summarizing health records
• Facilitating billing—such as checking insurance benefits, completing prior authorizations, and submitting claims

Clinical use

AI tools offer an array of options to augment and extend care beyond the traditional therapy visit.

Early detection

AI has the potential to aid in the early detection of individuals who may be at risk for developing mental health concerns. AI does this by noticing patterns from vast amounts of data; for example, AI could be used to analyze a person’s entire medical record to help identify those who might benefit from early intervention.

Clinical support

AI is being incorporated into many types of clinical support tools. Digital therapeutics, which are evidence-based, clinically validated software programs, are one category of digital tools that show considerable promise to augment care. For example, a clinician could treat a patient for one presenting concern, such as anxiety, and also could order the patient a digital therapeutic for use at home that could support their work (e.g., facilitate the practice of CBT-based skills to augment anxiety treatment) or to target other symptoms, like co-occurring insomnia.

Additionally, AI-enabled wearable devices can monitor symptoms and provide feedback both to the patient and clinician about a patient’s functioning. This category of devices offers opportunities to enhance real-time symptom monitoring, prompt the use of relevant therapeutic skills in the moment, and facilitate outcomes assessment.

Training use

AI has the potential to enhance many of the methods that are used to train clinicians. For example, many graduate students begin to learn how to provide therapy through engaging with case studies and mock patients. However, AI could be used to increase the exposure of clinicians-in-training to a much greater diversity of patients.

Additionally, AI can be used to monitor treatment to give clinicians real-time insights about risk or other important themes that emerge. It can also monitor adherence to a given treatment protocol and make suggestions for how to improve or tailor interventions.

While using AI in mental health care has considerable potential, there also are many ethical considerations that must be navigated to harness AI responsibly. Some of these considerations include mitigating algorithmic bias, obtaining informed consent from patients, and safely handling and protecting sensitive data. Given these ethical considerations, governments, industry experts, and patient advocate groups across the world are working to develop legal and regulatory frameworks to provide appropriate guardrails for the development and use of AI.

International landscape

There is a growing call for a global approach to regulating AI systems. Both the United Nationsopens in new window and the World Health Organizationopens in new window (WHO) have issued guidance on developing and maintaining responsible AI systems.

The WHO has identified the following “Key AI Principles” as foundational pillars that guide the use of AI in health care:

• Protect autonomy
• Promote human well-being, human safety, and the public interest
• Ensure transparency, explainability, and intelligibility
• Foster responsibility and accountability
• Ensure inclusiveness and equity
• Promote AI that is responsive and sustainable

European Union

In 2024, the European Union (EU) passed landmark AI legislation that applies to developers based in the EU, as well as products that are marketed to EU citizens.

The EU AI Actopens in new window takes a risk-based approach to regulating AI. It classifies AI systems from minimal risk to unacceptable risk and then regulates classes in increasingly stringent ways. A phased implementation of the provisions began in August 2024 and continues through August 2027.

United States landscape

In 2023, the Biden Administration released an Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. 

This was then followed by the release of a Blueprint for an AI Bill of Rights, which focused on five key principles:

1. Safe and effective systems: AI systems should be developed proactively to identify and mitigate risks and unintended consequences
2. Algorithmic discrimination protections: AI systems should be developed and used in equitable ways that do not unfairly treat groups of people
3. Data privacy: AI systems should be developed with safeguarding data privacy as a default and that seek user permission regarding how data is used
4. Notice and explanation: Explanations should be provided in clear and simple language explaining when and how AI systems are being used
5. Human alternatives, consideration, and fallback: People should be able to opt-out of AI systems and be provided with a human alternative

In 2025, upon taking office, the Trump Administration revoked the Biden Administration’s Executive Order on AI. The Trump Administration has issued a new executive orderopens in new window that focuses on removing barriers to innovation and positioning the U.S. as a global leader in AI. 

Federal agency oversight

While, presently, there is no overarching U.S. federal legislation of AI, key federal agencies have oversight over aspects of AI use.

Food and Drug Administration (FDA)

The FDA regulates medical devices that are enabled by AI and machine learning, as well as software as a medical device (SaMD). The Digital Center of Excellenceopens in new window within FDA was established with the goal of fostering digital health innovation. APA collaborates with FDA through participation in the Network of Experts program and as an advisor in the Total Product Lifecycle Advisoryopens in new window program to help support developers in bringing high quality digital mental health tools to market.

Federal Trade Commission (FTC)

The FTC protects the public from businesses that use AI to mislead or defraud consumers. For example, the FTC has been involved in investigating companies that have allegedly misused data, such as using sensitive customer data to train AI models without customers’ expressed consent.

There has been a significant expansion in the availability of AI-enabled tools that are marketed to mental health professionals. Given this, it is important that psychologists develop a process to evaluate any tools that they are considering for incorporation into their professional practice.

Mental health professionals’ needs may vary depending on the nature of their work and setting. Some key considerations include a review of the company and clinical leadership, tool functionality, research on the safety and effectiveness of the tool, compliance with HIPAA and other relevant data privacy regulations, data security, privacy practices, and terms of service.

APA has developed an evaluation guide that elaborates on these key considerations to assist psychologists as they review tools for their professional use.

As AI continues its rapid evolution, it will continue to transform the way that health care is practiced.

To stay up to date on innovations in mental health technology, visit our monthly “On The Horizon” column.

To view answers to common questions about the intersection between technology and psychological practice, visit OHCI’s monthly “Tech Talk with the Experts” series and consider submitting your own question.