The technology companies behind artificial intelligence-based note-taking software are sold to therapists as time-saving management tools, but quietly include regulations in their terms so that they can sell and manipulate patient care records to train other AI applications.
Providers outsource standard progress notes to automated software that summarizes session recordings and transcripts, which could unconsciously provide patient sensitive health information as data feed to the multi-billion dollar AI therapy industry.
“Most therapists don't exactly like writing progress notes. What happens if Cerapro writes them for you?” AI software companies charge mental health providers $60 per month for unlimited AI note-taking, including robotic diagnostic assistance and treatment plans.
According to Therapro's terms and conditions, when a provider signs up, it offers a “non-exclusive, transferable, assignable, perpetual, royalty-free, global license” for patients' anonymous treatment sessions.
This includes “not limited to” the right to train “artificial intelligence programs” developed by the company, including in cooperation with third-party contractors. Therapro reserves the right to “store, access and manipulate” identified patient data with third parties and deny any responsibility for how those contractors behave.
Research shows that “deprecated” anonymized health data (which allows tech companies to circumvent privacy laws) has become virtually obsolete thanks to advances in machine learning. The AI model successfully re-identified individuals from depersonalisation data with an accuracy of up to 85%.
“Our goal is to give therapists AI and not replace them, but to use de-consented data to enhance services, to constantly provide mission services, to help therapists better help patients heal,” Serapos wrote in a statement. tHe's a lever.
The founder of Freed says he created his AI Medical Scribe tool after years of watching his therapist wife write down a chart of patients late at night. The Software Terms and Conditions include the broad license and “right to sublicense, restock, perform, use, store, publish, perform, display, distribute, and more.
Other Scribe software companies hedge a wide range of terms, committing to protecting consumer data from using it to train other AIs.
BluePrint.ai tells therapist to “focus” [their] Client” and “Leave the Documentation to Us.” The terms and conditions grant unsolicited access to the blueprint “process, modify, reproduce, create, and create derivative works, display and disclose” and “use and share for purposes permitted under applicable law.”
Blueprint said lever The therapist always reserves the right to delete recordings and notes, and its session data is not used to train AI. “We are committed to better outline this in our Terms of Use,” founder Danny Fried wrote. “Blueprints exist to assist therapists and do not replace them.”
Meanwhile, SimplePractice is a software company that says it is “committed to the future of human-centered behavioral health” — pointing out that it currently does not store patient session data. . . However, once the opt-out option is created, plan it. The company said, “We will limit contractual restrictions… [patient] Data for training your own AI. ”
All of this is because one of the country's deepest out-of-pocket medical institutions dropped a massive investment in the development of AI-powered “robot” therapies. The Los Angeles-based hospital system Cedars-Sinai Medical Center recently created a fully automated “therapist” called Xaia.
VRX Health, a for-profit “digital therapy” company that holds an exclusive commercial license to XAIA, requires that users waive their rights to trial trial.
AI startups, including ChatGpt Creator Openai, are promoting automated chatbot technology for therapeutic applications, claiming that the software is perfect for counseling as it displays “human-like sensitivity.” The AI-powered healthcare startup raised $3.9 billion last year, including $1.4 billion for mental health. tHe is the Los Angeles Times Report.
Last week, the Federal Trade Commission launched a formal investigation into the safety of companion chatbot technology in light of the death of a 14-year-old Florida boy due to suicide after forming an abusive relationship with an AI bot. These relationships are becoming more and more common among teenagers. A July survey found that nearly three children, ages 13 to 18, used chatbots to date.
Dr. Beyerright, senior director of healthcare innovation at the American Psychological Association, testified before Congress earlier this month on the rise of what is called “digital therapy.” She said that therapy chatbots can “provide care to those who may otherwise receive nothing. These tools are most effective and safe when used for exchange rather than for care provided by qualified professionals, and ensure the body of the human in the loop.”
Freed CEO did not respond to requests for comment tHe is a lever We received responses from the company's “support bots” to confirm the company's policies regarding the dismantling of identification data.
