Concerns are rising among therapists regarding corporate platforms recording therapy sessions to train artificial intelligence models. While AI tools from companies like Blueprint offer benefits such as summarizing sessions and providing real-time prompts, experts like Vaile Wright of the American Psychological Association emphasize that AI cannot replace the vital human connection and clinical judgment in therapy. Colorado is now implementing new regulations for AI use by mental health professionals and insurance companies, and Utah is piloting an AI chatbot from Legion Health for prescription renewals, requiring a 98% physician approval rate for the first 250 prescriptions.
In AI development, Google DeepMind researcher Mostafa Dehghani notes that AI models are increasingly improving themselves, with new generations built heavily on previous ones, moving towards full automation. This recursive self-improvement is happening rapidly, making compute power the primary limitation. However, AI pioneer Sarah Douglas, a philosophy alum, points out that current large language models like ChatGPT are primarily statistical correlations and do not possess a true understanding of meaning or consciousness, a concept she has explored since the 1960s.
On the corporate front, AI company Anthropic, known for its chatbot Claude, has formed a political action committee, AnthroPAC, funded by employees. This move coincides with Anthropic's legal dispute with the White House over copyright infringement. Meanwhile, law firms like Brothers Smith LLP are actively training staff to leverage AI for improved attorney performance, aiming for increased speed, effectiveness, and cost-efficiency. AI expert witnesses are also becoming crucial in legal cases, helping courts understand complex AI concepts and analyze systems for biases and compliance.
Apple, after a slower start in the generative AI race, is now working to catch up, with former insiders suggesting its privacy-first approach may have cost it an early lead. While Apple previously excelled with devices and Siri, rivals focused on data-driven AI development. The company is now partnering with Google's Gemini to enhance Siri and plans to run more AI processes locally on devices. Despite past challenges, Apple is expected to remain a strong competitor in the evolving AI space.
Key Takeaways
- Therapists express concerns about corporate platforms recording sessions to train AI, citing privacy issues and the irreplaceable nature of human connection.
- AI tools assist therapists with tasks like summarizing notes but cannot replace human connection or clinical judgment, as stated by experts from the American Psychological Association.
- Colorado is implementing new regulations for AI use in mental health, and Utah is piloting an AI chatbot for prescription renewals, requiring a 98% physician approval rate initially.
- Google DeepMind researcher Mostafa Dehghani states AI models are recursively improving themselves, with new generations built on previous ones, making compute power the main limitation.
- AI pioneer Sarah Douglas highlights that large language models like ChatGPT are statistical correlations and lack a true understanding of meaning or consciousness.
- Anthropic, developer of the chatbot Claude, has formed AnthroPAC, a political action committee, amidst a copyright infringement dispute with the White House.
- Law firms, such as Brothers Smith LLP, are training staff to use AI for increased efficiency, speed, productivity, and cost-effectiveness in legal work.
- AI expert witnesses are essential in legal cases involving technology, intellectual property, and product liability, clarifying complex AI concepts for courts.
- Apple is partnering with Google's Gemini to improve Siri, aiming to catch up in the generative AI race after its privacy-first approach may have caused a slow start.
- Sharing personal medical records with AI chatbots like Claude and Perplexity for health advice carries significant data privacy risks, despite one-third of US adults seeking AI health advice.
Therapy sessions recorded to train AI models
Therapists are concerned about corporate platforms recording therapy sessions to train artificial intelligence. While some AI tools aim to replace therapists, experts like Vaile Wright of the American Psychological Association state AI cannot replace the human connection and clinical judgment vital to therapy. Companies like Blueprint offer AI tools that summarize sessions and provide real-time prompts for therapists, claiming strict data security. However, past issues with platforms like BetterHelp sharing data have raised privacy concerns among therapists.
AI tools help therapists but can't replace human connection
Artificial intelligence is increasingly used in mental health, with tools helping therapists with tasks like summarizing notes. Dr. Afarin Rajaei notes that while AI can be helpful, it should not replace therapy itself, especially for companionship. Colorado is implementing new regulations to guide the use of AI by mental health professionals and insurance companies. Experts emphasize that the human connection and presence in therapy are irreplaceable by AI, despite some optimism about AI's potential.
Law firms train staff on AI for efficiency and growth
Brothers Smith LLP is focusing on training strategies to help law firms effectively use AI. Partner C. Joseph Doherty highlights how AI can improve attorney performance by increasing speed, effectiveness, productivity, and cost-efficiency. This approach aims to accelerate the adoption of AI within law firms, leading to significant time savings and revenue growth through cross-functional collaboration.
AI company Anthropic forms PAC amid White House dispute
AI company Anthropic, known for its chatbot Claude, has formed a political action committee called AnthroPAC. This move occurs as Anthropic is in a legal dispute with the White House over copyright infringement. The PAC is funded by employees and aims to support candidates aligned with the company's interests. The timing of AnthroPAC's creation, alongside increased political focus on AI's impact on elections, has drawn attention.
Google DeepMind researcher: AI models are improving themselves
A Google DeepMind researcher, Mostafa Dehghani, states that AI models are increasingly improving themselves, a concept once considered science fiction. He explains that new generations of models are heavily built using previous ones, moving towards full automation. While not yet fully autonomous, this recursive self-improvement is happening rapidly and is visible in current AI products. Dehghani believes removing the human bottleneck in model improvement will lead to significant advancements, with compute power becoming the main limitation.
Philosophy alum Sarah Douglas programs AI to understand meaning
AI pioneer Sarah Douglas, a philosophy alum, focuses on programming computers to understand meaning, knowledge, and consciousness. She notes that current large language models like ChatGPT are statistical correlations, not truly thinking entities, and lack a concept of meaning. Douglas began her work in the 1960s, grappling with how to imbue computers with meaning. Her philosophical background at UC Berkeley helped her explore these complex questions in AI development.
Utah pilots AI for prescription renewals
Utah is testing a pilot program allowing an AI chatbot from Legion Health to renew drug prescriptions without immediate doctor approval. The program requires users to opt in, and the first 250 prescriptions will be monitored by a physician, needing a 98% approval rate to proceed without oversight. While aiming to expand access to mental health services, experts question the reliability of using chatbots for prescription renewals. This follows a previous AI prescription pilot program in Utah with Doctronic.
AI expert witnesses crucial in tech-related legal cases
Artificial intelligence expert witnesses are becoming essential in legal cases involving technology, intellectual property, and product liability. These experts, with backgrounds in computer science and data analytics, help clarify complex AI concepts like machine learning and natural language processing for courts. They analyze AI systems for biases, errors, and compliance with regulations, providing testimony on issues ranging from autonomous vehicles to AI-driven medical devices. Their role is vital in bridging the gap between technical AI and legal requirements.
Testing AI chatbots for health advice with personal medical records
The article explores the use of AI chatbots like Claude and Perplexity for health advice, with the author testing them using personal medical records. While one-third of US adults have sought AI for health advice, uploading sensitive medical data carries significant risks. The author expresses nervousness about sharing personal health information with chatbots, highlighting the need for caution and awareness of potential data privacy issues.
Apple's AI strategy: Catching up after a slow start
Former Apple insiders suggest the company's privacy-first approach may have caused it to lose a significant lead in the generative AI race. While Apple previously excelled in consumer devices and Siri, its rivals focused on data-driven AI development. Apple is now partnering with Google's Gemini to improve Siri and plans to run more AI on devices locally. Despite past challenges, insiders believe Apple can still compete effectively in the evolving AI landscape.
Sources
- Yes, Therapy Sessions Are Being Used to Train AI
- AI in therapy: Professionals weigh benefits and limitations
- Accelerate Effective, Efficient AI Law Firm Use with Concrete Training Strategies
- AI Giant Anthropic Files to Launch 'AnthroPAC' Amid Clash With Trump Administration
- Recursive Self-Improvement Of AI Models Is No Longer Sci-Fi: Google DeepMind Researcher
- Philosophy alum Sarah Douglas on her lifelong effort to program
- Utah Is Giving Dr. AI the Power to Renew Drug Prescriptions
- Artificial Intelligence Expert Witness: AI Analyst Scott Steinberg
- Can AI Be My Health Adviser? I Tested Claude and Perplexity With My Real Medical Records to See
- Apple at 50: The iPhone maker 'blew a 5-year lead' on AI, but former insiders say it can still win
Comments
Please log in to post a comment.