AI News Brief for Wednesday, May 13, 2026

  • Can You Trust AI with Your Retirement Planning? An MIT finance professor puts ChatGPT-5.2 through its paces on retirement guidance, and finds real promise alongside real limits worth knowing.

  • A Free AI Tool That Predicts Your Long-Term Care Needs. Waterlily uses your health history and family data to project personalized long-term care costs, tackling one of retirement's biggest financial unknowns.

  • Google's Fitbit Air Brings Affordable AI Health Coach to Your Wrist. Google's new screenless wearable pairs with an AI coach that tracks sleep, heart rate, and a-fib, but comes with privacy trade-offs worth understanding.

  • ChatGPT Health Launches a Private Space for Your Most Personal Questions. OpenAI opens a dedicated, privacy-protected health experience in ChatGPT, where 230 million people a week are already asking health questions.

Can You Trust AI with Your Retirement Planning?

Financial & Retirement
Source: MIT Sloan
https://mitsloan.mit.edu/ideas-made-to-matter/want-to-use-ai-to-plan-your-retirement-heres-how-to-proceed

In a nutshell

MIT Sloan finance professor Andrew Lo finds that ChatGPT-5.2 has improved significantly in personalizing retirement guidance and demonstrating emotional intelligence, but still lacks legal fiduciary responsibility, struggles with precise tax math, and cannot navigate regulatory nuance. Lo advises users to treat AI as a collaborative partner by prompting it to challenge their reasoning, state its assumptions, and cross-check conclusions across multiple platforms. Until accountability frameworks exist, individuals must stay informed and engaged to protect their own financial interests.

Why it's important

Directly addresses whether AI can be trusted for retirement financial decisions, a central concern for adults 50+, and provides specific, actionable steps for using it safely and skeptically.

Key takeaways

  • ChatGPT-5.2 has improved meaningfully in personalizing retirement advice. It can now tailor guidance to individual situations and, when asked about a major financial loss, responds with empathy before offering strategy. Professor Lo calls the shift significant.

  • AI still carries no legal fiduciary duty. If an AI tool gives you bad financial advice, there is no recourse. Lo is clear: treat it as a collaborative partner, not an authority.

  • Precise tax math and regulatory nuance remain weak spots. Large language models work by predicting likely responses based on training data, not by running strict arithmetic. Do not rely on AI for exact tax calculations.

  • Data security is an open question. Sensitive financial details entered into an AI platform may be used to train future versions of that model. Lo's blunt assessment: "We have no idea what happens to the data."

  • Lo offers five practical strategies for getting value from AI in retirement planning: ask it why you might be wrong, require it to state its assumptions, cross-check conclusions across multiple AI platforms, use it to teach you finance basics, and refine your prompts over time.

  • The bottom line, in Lo's framing: the newest AI models are genuinely useful retirement planning partners for informed, engaged users who verify what they hear. The era of AI as an oracle has not arrived yet.

Your thoughts?

When it comes to retirement planning, have you tried using an AI tool for guidance, and what was your experience?

A Free AI Tool That Predicts Your Long-Term Care Needs

Financial & Retirement
Source: Kiplinger
https://www.kiplinger.com/retirement/retirement-planning/i-tried-a-new-ai-tool-to-answer-one-of-the-hardest-retirement-questions-we-all-face

In a nutshell

A Kiplinger reviewer tested Waterlily, an AI platform that estimates personalized long-term care needs and costs using health data, family history, and a database of over 500 million data points from 50,000 families. The tool generates individualized care projections and connects users with financial strategies for funding future care needs.

Why it's important

Long-term care costs are among the biggest financial unknowns in retirement; a tool that personalizes predictions based on individual health history fills a real planning gap for adults 50+.

Key takeaways

  • Waterlily is a free AI-powered platform that draws on health data, family history, and a database of over 500 million data points from 50,000 families to project your personal long-term care needs and likely costs.

  • The tool generates individualized projections rather than broad national averages, which is what makes it genuinely useful. Most retirement planning guides quote median care costs; Waterlily estimates what care might cost for someone with your specific profile.

  • Long-term care remains one of the biggest financial wildcards in retirement, with costs shaped by health history, genetics, and geography. Most people underestimate both the likelihood they will need care and the price.

  • Beyond prediction, Waterlily connects users with financial strategies for funding future care, moving the tool from interesting data point to practical planning aid.

  • The review came from a veteran financial journalist at Kiplinger, a publication with a long record of credible, editorially independent personal finance coverage.

  • Like any AI tool working with personal health and financial data, Waterlily works best as a conversation starter with a qualified financial adviser, not a final answer on its own.

Your thoughts?

Have you used AI to predict long-term care in your retirement planning? What was your experience?

Google's Fitbit Air Brings Affordable AI Health Coach to Your Wrist

Health & Medicine
Source: CNN
https://www.cnn.com/2026/05/07/tech/google-ai-health-fitbit

In a nutshell

Google launched the Fitbit Air, a $99.99 screenless wearable, alongside the new Google Health app and an AI Health Coach powered by Gemini. This product offers personalized fitness plans, health record summaries, and conversational answers to health questions. The service is available to Apple users, marking a shift toward cross-platform AI health management. Note: Source is reporting on a Google commercial product launch; Google has a commercial interest in this topic.

Why it's important

An affordable AI-powered wearable with a personal AI coach lowers the barrier for older adults to manage their own health actively.

Key takeaways

  • Google launched the Fitbit Air ($99.99), a slim screenless wristband, alongside a redesigned Google Health app (replacing the Fitbit app starting May 19). The app includes an AI Health Coach powered by Google's Gemini models.

  • The AI Health Coach can summarize health records you choose to share, create personalized weekly fitness plans, and answer conversational questions about your fitness and medical data. It can also flag trends, for example noticing that a few extra hours of sleep improved your recovery metrics.

  • Google is opening the AI coach to Apple Watch, Oura ring, and Garmin users, not just Fitbit owners. The goal is to become the AI health layer across whatever hardware you already use.

  • Health institutions including the Mayo Clinic and Duke University School of Medicine caution that AI health tools can be inaccurate, lack the ability to run tests, and cannot replicate clinical reasoning. Google says the product is not intended for medical use or to replace doctors.

  • An important privacy note: the Fitbit Air and Google Health app are not HIPAA compliant for personal use. HIPAA protections apply only when the device is provided through a doctor or employer's structured program. Google has a well-documented history of using personal data to inform advertising.

  • Many of the most useful features require a separate Google Health Premium subscription, on top of any other Google plans you may already pay for.

Your thoughts?

Would you consider wearing a $100 AI-powered health tracker to keep a closer eye on your health?

ChatGPT Health Launches a Private Space for Your Most Personal Questions

Health & Medicine
Source: OpenAI
https://openai.com/index/introducing-chatgpt-health/

In a nutshell

OpenAI has launched ChatGPT Health, a dedicated service designed to help users navigate health questions and access personalized health information. The product represents a significant expansion of AI into consumer healthcare guidance. Note: Source has a commercial interest in this topic.

Why it's important

Older adults who are among the heaviest users of healthcare can now access an AI health information service; understanding its capabilities, limitations, and data practices is essential before relying on it.

Key takeaways

  • ChatGPT Health is a separate, private space within ChatGPT built specifically for health conversations. It has its own memories and storage, and what you share there stays contained. Health conversations are not used to train OpenAI's models. However, the app is not HIPAA compliant unless it’s provided by your doctor or employer, which means your health data could be used for other purposes.

  • OpenAI reports that over 230 million people globally already turn to ChatGPT with health questions every week, making health one of the most common uses of the platform. ChatGPT Health adds a dedicated layer of privacy and personalization on top of that.

  • Users can securely connect medical records and wellness apps (including Apple Health, Function, and MyFitnessPal) so the AI can help make sense of recent test results, prepare questions for a doctor appointment, or compare insurance options based on your actual health patterns.

  • The service is clear about what it is not: it is not for diagnosis or treatment, and it is not a replacement for doctors or clinicians. Its role is to help you feel more informed and prepared.

  • Access is rolling out gradually, starting with a small group of Free, Go, Plus, and Pro plan users outside the European Economic Area, Switzerland, and the UK. Medical record integrations are US-only, and Apple Health requires iOS.

  • See also the related article in the neoClarity community: https://community.neoclarity.ai/c/articles-and-research/openai-unveils-chatgpt-health

Your thoughts?

If you could securely connect your medical records to an AI that helped you prepare for doctor visits and understand test results, would you be willing to share sensitive health information with the app even if it's not HIPAA compliant?





This brief was prepared by Martin Danner (a curious human) using Claude Cowork as a research assistant.