What I Tell AI, What I Keep To Myself, And How It Watches My Back

A contributor to Sixty and Me says she used ChatGPT to identify local care and insurance-support services for her sister with Parkinson’s, entering a general description of the situation and a ZIP code rather than names or account details. Her account highlights both the practical use of AI for finding resources and the need to protect privacy and verify its suggestions.

The writer recounts that her eldest sister called in a panic because their middle sister, who is in her 70s and lives with Parkinson’s, was struggling to obtain in-home care. The siblings lived thousands of miles apart, and the writer said she felt unable to help directly. She asked ChatGPT to suggest local advocacy and insurance resources, providing the sister’s condition, approximate age and ZIP code.

According to the article, ChatGPT returned a list that included a local Area Agency on Aging, a free Medicare counseling program, the Parkinson’s Foundation helpline and a legal aid organization. The writer says she began calling, left messages and spoke with one person who reassured her that the organization could help. The account does not independently confirm the accuracy or availability of every listing.

The writer says she does not provide full names, birthdates, doctors’ names, insurance or Medicare numbers, or medical documents showing identifying information. She describes using AI to explain medical terms, prepare questions for appointments and find organizations, while treating the answers as research rather than medical advice. For supplement questions, she says she may enter product details and an expiration date, but acknowledges AI cannot guarantee an expired product is safe or effective.

Using AI Without Sharing an Identity

The account offers a concrete example of how a conversational AI tool can help someone begin a search for local services when family members are distant. The writer’s approach distinguishes between information that may help narrow a search, such as a condition and location, and identifiers that could expose a person’s identity or accounts.

That distinction matters because health-related questions can involve sensitive personal information. The article is an individual’s experience, not a test of ChatGPT’s privacy protections or a guarantee that its answers are correct. Its practical takeaway is to limit what is entered, check contact details with official sources and take health decisions to qualified professionals.

From a Care Search to Everyday Questions

The article began with a request for help finding advocacy and insurance support for a relative with Parkinson’s. The writer says her sister had been unable to secure needed in-home care and believed she had paid for such care through insurance during her career. The report does not specify the policy, the nature of the dispute or what services the sister ultimately received.

From that episode, the writer describes using AI for other everyday research, including clarifying medical language and preparing for appointments. She advises describing a person’s role rather than identity and says that, when asking about symptoms, users can share plain-language descriptions, duration and relevant background without submitting identifying records. She also recommends checking details on official websites, since AI can get specifics wrong.

““I’m not getting anywhere near AI. I don’t trust it.””

— The writer, recounting her sister’s response

Accuracy and Privacy Questions Remain

The source is a first-person account and does not include an independent review of the recommendations ChatGPT supplied. It is unclear whether all listed phone numbers and services were current, whether the sister obtained in-home care, or whether her insurance concern was resolved. The writer reports reaching one live contact but gives no name or outcome for that conversation.

The article also does not assess how the AI service handles prompts or explain the privacy settings in effect. Avoiding names and account numbers can reduce the amount of identifying information shared, but the report does not establish that any particular prompt is private or that a ZIP code cannot contribute to identification. Its description of supplement research likewise does not establish that AI can determine product safety.

Verify Resources Before Acting

The writer says she followed up on the suggestions by calling organizations and checking details. Readers using AI to locate care or support can confirm names, phone numbers, eligibility rules and services through official organization websites or direct contact. The source gives no scheduled follow-up or reported resolution of the sister’s care needs.

For health concerns, the article’s stated approach is to use AI to prepare questions and discuss information with a care team before making changes. It also advises seeking urgent help for serious symptoms rather than relying on an AI response; the supplied article excerpt ends while beginning an example of urgent symptoms, so it does not provide a complete list.

Key Questions

What did the writer ask ChatGPT to find?

She asked for local help for her sister, who has Parkinson’s and needed an advocate for care and help sorting out insurance. She says the response listed an Area Agency on Aging, Medicare counseling, a Parkinson’s Foundation helpline and legal aid.

What personal information did she share?

She says she supplied the sister’s condition, approximate age and ZIP code. She says she did not enter the sister’s name, birthdate, doctors’ names or insurance numbers.

Does the article show that ChatGPT’s resource list was fully accurate?

No. It reports the writer’s experience, not an independent check of every listing. She advises verifying phone numbers and details on official websites.

Does the writer recommend using AI as a doctor?

No. She calls AI a research partner and says its information should not replace a clinician’s judgment or be used alone to make health decisions.

Source: rss