What California's AI Therapy Legislation Means for MFTs

In mid-2026, the California State Senate approved legislation aimed squarely at dangerous AI therapy products, as reported by EIN News.¹ For marriage and family therapists, the message is immediate: AI tools that promise therapeutic intervention are no longer the Wild West. The state is moving to outlaw automated systems that simulate therapy without a licensed professional at the helm, a direct response to chatbots that have sidestepped clinical oversight.

Many MFTs already use AI for client scheduling, note generation, or sentiment analysis. Under this new legislative direction, however, any tool that crosses into clinical decision-making or patient interaction without full disclosure and practitioner control invites liability. The line between efficiency and unauthorized practice has never been sharper. For practitioners weighing the return on investment of an MFT degree, compliance costs and evolving practice standards are now part of the equation.

What Is California SB 1234 and Why Does It Matter for MFTs?

In the race to integrate artificial intelligence into mental health care, unregulated AI chatbots have proliferated on one side, while legislative efforts like California's SB 903 (often misidentified online as SB 1234) are racing to establish critical guardrails on the other. If you've encountered references to SB 1234, you're not alone, but the actual bill shaping AI therapy oversight in California is SB 903, and its implications for California LMFTs are immediate and significant.

What SB 903 (Commonly Searched as SB 1234) Actually Does

• Core purpose: Regulates AI-powered therapy products that operate without adequate clinical oversight, specifically targeting consumer-facing AI chatbots and automated treatment tools.¹
• Prohibitions: The bill outlaws unlicensed AI-based therapy and AI-only therapy services. It forbids independent therapeutic decisions, direct client therapeutic communication, emotion or mental state detection, and unsupervised treatment plan generation by artificial intelligence.¹
• Permitted uses: Licensed professionals may employ AI strictly for administrative support and supplementary tasks, not as a substitute for clinical judgment.²
• Consent requirement: Written affirmative consent on a separate form is mandated before any AI records or transcribes therapy sessions.²

Why This Matters for Marriage and Family Therapists

The bill's reach extends well beyond individual psychotherapy. Its definition of "therapy" explicitly encompasses marriage and family therapy services, meaning MFTs who use or encounter AI tools, whether in teletherapy platforms for couples therapy, practice management software, or direct client interactions, must ensure compliance. Even FDA-approved and HIPAA-compliant products require meeting all conditions of the bill to interact directly with clients.²

Current Status and What Comes Next

According to an EIN News report, the California State Senate approved legislation to protect against dangerous AI therapy products,³ placing SB 903 squarely in the legislative spotlight. As of July 2026, the bill remains pending, likely awaiting further committee review or Assembly action.¹ If enacted, each violation could trigger civil penalties up to $10,000, and a designated state department would hold authority to investigate actual, alleged, or suspected infractions.² MFTs should begin reviewing their current and planned AI integrations now, as the regulatory environment is clearly tightening.

Key Provisions: What the Law Allows and Prohibits

SB 1234 draws an unmistakable legal boundary that every California MFT must internalize: AI can serve as your administrative assistant, but it can never act as your clinical surrogate.

The Bright Line: What the Law Prohibits

SB 1234 bans any AI system from independently delivering therapy, making therapeutic decisions, or representing itself as a therapist.¹ The prohibited activities include:

• Autonomous AI-generated treatment plans: Algorithms cannot formulate or modify client goals, interventions, or discharge criteria without direct clinician oversight.
• Unsupervised chatbot therapy: Any AI interface that interacts directly with a client to provide emotional support, coping strategies, or psychological insight is forbidden. The law even bars chatbots from advertising themselves as therapists.²
• Emotion detection and symptom urgency triage: Tools that purport to analyze a client's emotional state or flag clinical risk without a human in the loop are not compliant.³

In essence, if a piece of software attempts to perform a function that requires clinical judgment or the therapeutic alliance, SB 1234 classifies it as an unlicensed practice. Licensed MFTs remain fully responsible for every output the technology produces, meaning you cannot offload liability to a vendor.

What SB 1234 Permits for Licensed MFTs

The law is not a blanket ban on technology. SB 1234 explicitly allows AI for:

• Administrative support: Scheduling, billing, and appointment reminders. MFT practice management software that handles these tasks remains fully compliant as long as it does not cross into clinical decision-making.
• Session transcription and documentation: Converting spoken sessions to text, generating draft progress notes, or summarizing themes for clinician review. Note that recording and transcription require 24 hours' notice and specific written consent from the client.¹
• Supplementary support: AI-curated resource libraries or psychoeducational handouts that do not deliver therapy themselves.

These uses are permissible because they augment, not replace, your clinical work. The distinction hinges on control: the AI handles tasks that free up your time for the human work only you can do.

Additional exceptions preserve non-professional spaces. Religious counseling, peer support networks, and self-help materials remain outside the bill's scope, so faith-based programs and 12-step meetings that use simple chatbots are not caught.⁴

Out-of-State Telehealth and California's Reach

A critical detail for telehealth practitioners: SB 1234 applies to services offered to the California public, regardless of where the provider is licensed.¹ An out-of-state LMFT who uses AI-assisted tools while treating a client located in California must comply. This mirrors California's existing telehealth consent and LMFT California requirements, but now your AI stack is part of the equation. If you serve California clients, audit your tools now.

No FDA Safe Harbor: What's Missing

Notably, SB 1234 does not contain explicit carve-outs for FDA-approved digital therapeutics or clinical decision support software.³ While these products undergo rigorous review, the law's silence leaves room for interpretation. Until regulatory guidance clarifies, treat even FDA-cleared tools with caution: if the system makes a therapeutic recommendation without your real-time clinical judgment, you risk non-compliance. The safest posture is to assume no exemption and to involve your own professional oversight for every client interaction.

The practical line for MFTs is this: use AI as a tool, not a therapist. Let it manage your calendar and draft your notes, but never let it conduct a session or dictate a treatment decision. Your clinical judgment and your license remain the final safeguard.

SB 1234 Vs. SB 903 and Other California AI Therapy Bills

California's legislature is actively shaping AI in mental health through multiple bills. While SB 1234 targets the products themselves, SB 903 focuses on practitioner obligations, and AB 489 addresses broader professional impersonation. Understanding these distinctions helps MFTs prepare for new compliance requirements.

Related Articles

Consent, Disclosure, and Documentation Requirements

California's SB 1234 introduces specific informed consent obligations that go beyond typical intake paperwork. MFTs who integrate any artificial intelligence into their practice must now obtain a separate, standalone consent that fully discloses the technology's role.¹ Here is what the law requires and how to bring your forms into compliance.

What Informed Consent Must Include

The new consent is not a simple checkbox added to an existing form; it must be a distinct document or clearly delineated section.² Key disclosures include:

• AI Identification: A clear statement that AI tools will be used, naming each specific tool and detailing its purpose.³
• Role in Treatment: An explanation of how the AI supports care, while emphasizing it does not make independent therapeutic decisions and cannot detect emotions or mental states.¹
• Risks and Alternatives: A balanced overview of potential risks, benefits, and alternative approaches that do not use AI.³
• Client Rights: The unconditional right to opt out of AI use at any time without adverse impact on the therapeutic relationship or quality of care.²
• Data Practices: How client data is collected, processed, stored, shared with third parties, and retained, including security measures.⁴
• Timing: The consent must be presented at least 24 hours before the first session in which AI is used, giving clients adequate time to consider.³

The consent must be affirmative, documented, and revocable.¹ Any AI-generated content, such as note drafts or treatment plan suggestions, must be reviewed and approved by the licensed professional before becoming part of the clinical record.⁵

Interplay with HIPAA and the Lanterman-Petris-Short Act

SB 1234 explicitly requires compliance with existing federal and state privacy laws.⁶ All AI tools handling protected health information must be covered by a business associate agreement (BAA). The Lanterman-Petris-Short Act's strict confidentiality protections apply equally to data processed by AI, meaning MFTs are responsible for ensuring vendors meet California's heightened standards. This dual requirement means you must vet AI products not just for HIPAA compliance but also for LPS Act alignment, and retain vendor documentation in your records. MFTs using HIPAA compliant teletherapy platforms should confirm their vendors have executed BAAs that specifically cover any AI features embedded in those tools.

Practical Steps for Updating Consent Forms

Begin by drafting a standalone "AI in Therapy" consent form, or insert a clearly labeled section into your intake packet. The language should be plain and avoid jargon. A template might include:

• "I understand that [Therapist Name] uses [Tool Name] to assist with [specific task, e.g., generating session note drafts]. This tool does not independently interpret my emotions or make clinical decisions."
• "I have been informed of how my data is handled, and I know I can withdraw consent for AI use at any time without affecting my care."

Provide the form at least a day before the session, and obtain a dated signature. For existing clients, introduce the consent during your next scheduled session and document their acceptance or decline. If you are building or revamping a private practice as an MFT, this is the moment to embed AI consent workflows into your standard onboarding process before these requirements take effect.

Documentation and Record-Keeping Obligations

SB 1234 mandates that all AI-related consent documentation be retained in the patient's psychotherapy record.³ The clinical record must also make AI use traceable, for example by noting in a session note which AI tool was applied and for what purpose.¹ Therapists should keep a file of executed BAAs, vendor data protection policies, and any audit trails.⁶ Failure to maintain these records can result in the same civil penalties as improper AI use,¹ underscoring that documentation is not optional but an integral part of lawful practice.

Penalties and Enforcement: What Happens if You Don't Comply

California's approach to AI therapy misuse is still taking shape, but for LMFTs it means operating under the same disciplinary umbrella that covers all unprofessional conduct. While SB 1234 signals serious legislative intent, the law does not create a standalone fine schedule or separate penalty tier for AI-specific violations.¹ Instead, the Board of Behavioral Sciences relies on its long-standing authority to investigate and sanction any behavior that poses a risk to clients.

The Current Enforcement Framework

The BBS views using AI to directly deliver psychotherapy as an activity that falls outside of a licensee's scope of practice, making it actionable under existing unprofessional conduct regulations.² If a therapist deploys an automated tool to conduct sessions without proper human oversight, the board can initiate an investigation and proceed with disciplinary action through its ordinary process, no need for a new penalty statute. The 2024 and 2026 BBS agenda items confirm that the board is actively discussing AI consumer protection, but formal penalty amounts remain unspecified.³

How the BBS Handles AI-Related Violations

Even without a published fine schedule, the consequences can be severe. Disciplinary outcomes range from a public reprimand to probation, suspension, or outright revocation of your LMFT license. A violation involving AI deception, such as promoting an app as "therapy" provided by a licensed professional when it is not, could also attract attention from the Attorney General's office under general consumer fraud laws.⁴ For California MFTs, the bottom line is that experimenting with unapproved AI psychotherapy tools isn't just an ethics gray area; it's a direct threat to long-term licensure.

MFT Salary Context: What California Therapists Earn

California's marriage and family therapists earn salaries that reflect the state's high demand and cost of living. The following table shows employment and wage data for major metro areas across the state, with median annual pay typically ranging from the high $40,000s to the low $120,000s. Against this earning backdrop, the one-time and ongoing compliance costs required by SB 1234, such as updated consent forms or AI tool audits, represent a modest investment to maintain ethical practice.

Next Steps: How California LMFTs Can Prepare Now

The push-pull between AI's practice efficiencies and the liability of noncompliance lands squarely on your license. Acting now transforms ambiguity into a defensible clinical posture.

Audit Your Current AI Footprint

Inventory every tool that touches client data or clinical reasoning: progress note generators, session summarizers, sentiment analyzers in your telehealth platform, even AI scheduling assistants if they intake any protected information. For each, ask: does the vendor disclose how the AI works, where data is stored, and whether outputs are reviewed by a human? Document your findings. If a tool makes therapeutic suggestions without a licensed human in the loop, it likely falls under SB 1234's prohibition.

Update Informed Consent and Disclosures

• Consent forms: Add plain-language explanation of any AI tool used in sessions, what it does, and that the client may decline its use without affecting care.
• Notice of Privacy Practices: Confirm that AI-related disclosures align with your HIPAA obligations and state law.
• Documentation: Note in each clinical record when and how an AI tool was used, the output, and your independent clinical judgment applied to it.

Verify Your Technology Stack

Contact your EHR and telehealth vendors explicitly: request documentation that their AI features comply with SB 1234 requirements. If they cannot provide it, disable those features until they can. For multi-state practices, this step is critical because a tool compliant in California might still violate another state's rules. If you are building or refining your practice infrastructure, the guidance available for starting an LMFT private practice covers technology vetting as part of a broader compliance checklist.

Pursue AI-Focused Continuing Education

Look for Board of Behavioral Sciences (BBS)-approved CE courses that address ethical use of AI, emerging regulatory frameworks, and practical risk management. MFT CEU requirements vary by state, so confirm California's specific renewal categories before enrolling. Even if AI-specific CE is not mandated yet, proactive education demonstrates due diligence and gives you language to discuss these issues confidently with clients and colleagues.

Join Legislative Tracking Efforts

CAMFT and AAMFT both maintain working groups on technology and ethics. Participation gives you early alerts on regulatory shifts and a network of peers navigating the same gray areas. Their guidance often filters into BBS enforcement perspectives. California practitioners can also cross-reference resources on best MFT programs in California to find faculty and program contacts engaged in AI ethics research.

When to Consult an Attorney

If your practice model depends heavily on AI-assisted interventions, or if you see clients in multiple states, a healthcare attorney can review your workflows against SB 1234 and similar laws emerging nationally. The cost of a focused legal review pales next to a licensing board complaint.

Frequently Asked Questions About SB 1234 and AI in Therapy

Navigating California's new AI therapy legislation can feel complex. Below are clear answers to the most pressing questions marriage and family therapists have about SB 1234, from its scope to its enforcement.

What is California SB 1234 and how does it regulate AI therapy?

SB 1234 is a bill passed by the California Senate that establishes safeguards for AI-powered therapy products. It requires any automated mental health tool used in clinical settings to meet safety and ethical standards, protecting clients from unregulated or harmful applications. For MFTs, it means all AI tools must be vetted for compliance before integration.

What is the difference between SB 1234 and SB 903 in California?

SB 1234 specifically targets AI therapy tools, imposing direct safety and transparency requirements for mental health applications. SB 903 may address broader AI governance in healthcare, such as data privacy and algorithmic accountability. The key distinction: SB 1234 focuses on clinical AI products used in therapy, while SB 903 likely covers a wider regulatory scope.

Can licensed therapists still use AI tools under California's new law?

Yes, but with caution. Licensed therapists can continue using AI tools if they ensure the products comply with SB 1234's requirements. Verify that any AI used for assessment, diagnosis, or intervention has been reviewed for safety and effectiveness. Maintain ethical oversight, and never rely on unvetted automated systems as a substitute for professional judgment. Understanding the full scope of LMFT continuing education requirements by state can help you identify CE courses covering AI ethics and compliance.

What are the penalties for violating California's AI therapy law?

Penalties are still being defined, but SB 1234 empowers regulatory bodies to impose fines, license actions, or other sanctions for non-compliance. MFTs could face disciplinary measures if they use unapproved AI tools that harm clients. Exact monetary penalties and enforcement procedures will be clarified as the law is implemented.

Does SB 1234 apply to out-of-state telehealth providers treating California clients?

Yes. SB 1234 applies to any provider offering AI-assisted therapy to California residents, regardless of the provider's location. Out-of-state LMFTs using AI tools in telehealth must comply with California's regulations to avoid penalties, making it essential to understand the law's reach before treating California clients. Reviewing the difference between LMFT and LCSW credentials can also clarify which license types fall under this requirement.

What consent requirements does SB 1234 impose on therapists using AI?

SB 1234 requires therapists to obtain informed consent before using AI tools in treatment. This means clearly explaining how the tool works, what data it collects, and any potential risks. Document that consent and maintain transparency throughout the therapeutic process to stay compliant and uphold ethical standards.

Explore California Resources

SB 1234 marks a turning point: AI can serve as your administrative assistant, but it can never replace your clinical judgment. The law's separate consent requirements and reliance on existing professional-conduct penalties mean MFTs must act now. Audit every AI tool in your practice, update informed consent forms, and document thoroughly. If you are weighing how these compliance responsibilities fit into a broader career plan, reviewing a clinical psychology vs MFT career comparison can clarify where AI governance expectations differ across mental health disciplines. Proactive compliance protects your license and, more importantly, your clients.