ChatGPT just started showing ads. Here's what therapists need to know

This week: ChatGPT introduces sponsored listings, changing how users find information; Google reviews vanishing from therapy practices nationwide; States race to regulate AI in therapy: Tennessee, Colorado, and 10+ states advance bills

ChatGPT introduces sponsored listings, changing how users find information

Over the past few weeks, OpenAI updated ChatGPT with new web-browsing and research features that allow the assistant to search websites, summarize information, and help users compare options. These tools are available first to Pro users, with Plus and Team tiers following. Separately, OpenAI has begun testing sponsored listings for Free and Go-tier users in the U.S., clearly labeled and visually separated from ChatGPT's responses. OpenAI states sponsored content does not influence the assistant's answers.

For users seeking therapy, ChatGPT can now gather information from websites and summarize available options, making it easier than ever to compare practitioners without opening a single browser tab. While users still make the final decision about which therapist to contact, the direction is clear: AI is rapidly becoming the first place people go for information, and it will increasingly replace traditional search as the primary way patients discover and evaluate providers. Practices with clear, accurate, and consistent online information will be the ones AI tools surface first.

๐Ÿ‘‰ Action: Ensure your website is clearly structured with specific service pages, credentials, insurance information, and contact details so any research tool can easily find accurate information. Make sure your Psychology Today, Google Business Profile, and website information are consistent across every platform. Write a comprehensive FAQ page answering common patient questions. Test how ChatGPT describes your practice by typing "Tell me about [your practice name] in [your city]" and correct any inaccuracies.

ChatGPT introduces sponsored listings, changing how users find information

Google reviews vanishing from therapy practices nationwide

Starting February 16, a massive spike in Google review disappearances hit businesses across every industry. By February 17, an unusually large number of businesses were reporting losses of 3 to 500 reviews overnight, with five-star reviews deleted at higher rates than negative ones. Google quietly updated its review policies the same week, tightening rules around incentivized and selectively solicited reviews.

For therapists, this hits differently than any other industry. As mental health professionals, you cannot and should not solicit reviews from clients. Period. Asking a client to leave a review can breach confidentiality, creates dual relationship concerns, and violates APA, NASW, and most state licensing board guidelines. HIPAA adds another layer: even acknowledging someone is your patient publicly is a violation. This means therapists have zero ability to "rebuild" lost reviews the way other businesses can. Every organic review is irreplaceable. For a practice with 30-50 reviews, losing even 3-5 can devastate your rating and patient acquisition.

๐Ÿ‘‰ Action: Log into your Google Business Profile immediately and screenshot your current review count and star rating. Do NOT ask clients for reviews, not now, not ever, this is an ethical and HIPAA issue, not just a Google policy. If you've lost reviews, report through Google Business Profile support with specifics. Set up notifications to catch further changes. Focus on what you control

Google reviews vanishing from therapy practices nationwide

States race to regulate AI in therapy: Tennessee, Colorado, and 10+ states advance bills

The week of February 20 brought a cascade of state legislative action on AI in mental health. Tennessee's Senate passed legislation prohibiting AI systems from representing themselves as qualified mental health professionals. Colorado introduced HB26-1195, making it an unfair trade practice for AI to imply its output equals licensed psychotherapy. Virginia's SB 796 (AI Chatbots and Minors Act) unanimously passed committee. Michigan introduced a kids chatbot safety bill. Over 300 AI-related bills are now tracked across U.S. state legislatures in 2026, with chatbot regulation as the dominant theme.

Illinois and Nevada already ban AI for behavioral health entirely. Ohio, Florida, and Pennsylvania have pending bills. For therapists, this creates both protection and obligation: AI legally cannot replace you in a growing number of states, but you must ensure your own practice's AI tools comply. The line between an administrative chatbot and a clinical interaction is where enforcement will focus.

๐Ÿ‘‰ Action: Search your state legislature's website for pending AI therapy or chatbot bills. Audit every AI tool that interacts with clients: website chatbots, scheduling assistants, intake automation, and follow-up systems. Ensure each clearly discloses it is not a human therapist. Update informed consent documents to include AI tool disclosures. Use these laws as a marketing advantage: emphasize that your practice provides human, licensed therapy.

States race to regulate AI in therapy: Tennessee, Colorado, and 10+ states advance bills

Google and AI now demand proof you're real: why grounding your entity matters more than ads

From February 15โ€“20, Google Search experienced increased ranking volatility. Google expanded AI Mode to 53 new languages, and John Mueller (Google Search Advocate) confirmed that force-indexing pages is discouraged. In addition, AI systems and search engines increasingly evaluate practices based on credibility indicators. Information such as practice name, address, phone number, license, photos, reviews, and website content contribute to determining whether a practice is active and legitimate.

Leads generated through paid advertising exist only while the campaign is active. Organic traffic, in contrast, comes from users who actively search for services and select providers based on relevance and trust. These users generally have higher conversion rates and are more likely to become long-term patients. Paid campaigns can provide short-term increases in leads, but a well-structured digital presence, including accurate business information, website quality, local SEO, and online reputation is essential for sustainable client acquisition.

๐Ÿ‘‰ Action: Audit your โ€œentity signalsโ€ by ensuring your practice name, address, and phone number are consistent across your website, Google Business Profile, Psychology Today, and other directories. Include your license number, NPI, and professional credentials on your website and in schema markup, and upload multiple authentic photos to your Google Business Profile. Before spending further on paid advertising, evaluate whether clients would still find your practice if all paid campaigns were paused; if not, prioritize organic channels such as email lists and directory optimization, which are channels you fully control.

Google and AI now demand proof you're real: why grounding your entity matters more than ads

OpenAI launches "Lockdown Mode" for ChatGPT Healthcare, retires older models

On February 13, OpenAI introduced Lockdown Mode across ChatGPT, ChatGPT Atlas, and Codex, with ChatGPT for Healthcare specifically included. Lockdown Mode restricts how ChatGPT interacts with external systems, combating prompt injection attacks where malicious actors trick AI into revealing sensitive data. When enabled, web browsing is limited to cached content, deep research and agent features are disabled, and file downloads are blocked. The same day, OpenAI retired GPT-4o and several legacy models, making GPT-5.2 the default.

If you use ChatGPT to draft clinical notes, research treatment approaches, or handle any practice management tasks, prompt injection attacks could potentially expose data. Lockdown Mode is available for Enterprise, Edu, and Healthcare plans, with consumer plans coming later. For therapists, the message is clear: if you're using AI tools with practice-related information, you need enterprise-grade security, not the free tier.

๐Ÿ‘‰ Action: Audit which ChatGPT plan your practice uses. If you're on a free or Plus plan handling practice-related work, consider upgrading to ChatGPT for Healthcare or Enterprise. Never paste patient information or session notes into any AI chatbot. Ensure all staff understand which tools are approved. Enable Lockdown Mode if you have access. Add AI security protocols to your HIPAA compliance documentation.

Quick wins for the week

  • Test how ChatGPT describes your practice by searching your name directly, and ensure your website and profiles have clear, consistent, factual information
  • Screenshot your Google Business Profile review count now to document any disappearances, and remember: never solicit reviews from clients due to ethical and HIPAA obligations
  • Search your state legislature's website for pending AI therapy regulation bills and review compliance requirements
  • Ground your entity: verify your name, address, phone, and credentials are identical across your website, Google Business Profile, Psychology Today, and all directories
  • Before spending more on Google Ads or Meta, ask: would patients still find me if I turned off paid ads today? If not, invest in your organic foundation first
  • Audit your ChatGPT usage: upgrade to a Healthcare or Enterprise plan if you handle any practice-related work, and never enter patient data

Final thoughts

This week made one thing undeniable: AI tools are becoming part of how people research and find services, including therapy. ChatGPT is introducing sponsored listings, expanding its research capabilities, and OpenAI is building healthcare-specific security features because the stakes around sensitive information are that high. Meanwhile, Google is deleting reviews at record rates, and therapists, unlike every other business, cannot simply ask for more because of the ethical and HIPAA obligations that define our profession. State legislatures are responding, with over 300 AI bills in play and a growing consensus that AI cannot legally replace a licensed therapist.

The through-line is trust. Your entity signals matter: consistent information across every platform, genuine credentials, authentic photos, and a foundation that proves you exist and serve real patients. No amount of paid advertising can substitute for this. The practices that will thrive are those that own their digital presence rather than renting it, and that understand the growing importance of having clear, accurate information everywhere patients and research tools look.

While you're providing therapy, we're monitoring policy changes, protecting your Google presence, and optimizing for AI search. You shouldn't need to become an SEO expert, compliance specialist, and tech strategist on top of being a therapist.