❓ Help Maaasahan mo ba ang AI about Health Advice?

Maasahan mo ba ang AI pagdating sa Health Advice? For example, kay ᑕᕼᗩTGᑭT or Gemini mo iaasa yung choice of food at daily routine mo. Gagawin mo lahat ng suggestions niya. Reliable ba para sayo ang ganon? I'm not talking about medical advice, only health and wellness advice lang.
 
Di boss. May na nakita Kasi Ako na Balita na nag ask Yung lalaki sa gpt na Anong magandang pamalit sa asin ang sagot ni gpt is "sodium bromide" ata yun. Yun sinunod nya ito ginagamit nya pangsagol sa lahat ng consume nya. After a week admit sya sa hospital muntik Ng ma deds.
 
Di boss. May na nakita Kasi Ako na Balita na nag ask Yung lalaki sa gpt na Anong magandang pamalit sa asin ang sagot ni gpt is "sodium bromide" ata yun. Yun sinunod nya ito ginagamit nya pangsagol sa lahat ng consume nya. After a week admit sya sa hospital muntik Ng ma deds.
depende kasi yan sa pagbabato ng prompt at model na gpt na gagamitin. Free version will give you general answer and siguro sa case nung lalake possible incomplete ang prompt na binigay nya.
 
usually naglalagay ng disclaimer ang sagot nila kaya overall ang effectivity at reliability nya ay naka depende sa user
 
Para sa’kin hindi siya auto‑reliable, depende sa situation. Ok yung AI pang-idea lang, parang ka‑chat na may suggestions, pero hindi yung tipong lahat ng kain mo at routine mo siya na magdidikta. Mas safe kung gawin mo lang siyang parang second opinion, tapos i‑double check mo pa rin sa ibang source or sa totoong doktor lalo na kung long term or may sakit ka na talaga
 
It can be reliable depende po sa gumagamit at set of prompts na meron ka.
Yung prompts talaga nandyan nakasalalay

Tip: Always put according to... based on...

Ex. Management of diarrhea according to... based on... latest Internal Med guidelines
from the Philippines please be specific

Latest Internal Medicince - pag ãdül† and concern
Pediatric - pag pedia

Derma - balat
Gastroenterology pag mga tiyan ang probs, etc.

Sa Google Gemini, medyo pipilitin mo sya na mag bigay ng output
 
It can be reliable depende po sa gumagamit at set of prompts na meron ka.
Yung prompts talaga nandyan nakasalalay

Tip: Always put according to... based on...

Ex. Management of diarrhea according to... based on... latest Internal Med guidelines
from the Philippines please be specific

Latest Internal Medicince - pag ãdül† and concern
Pediatric - pag pedia

Derma - balat
Gastroenterology pag mga tiyan ang probs, etc.

Sa Google Gemini, medyo pipilitin mo sya na mag bigay ng output
Totally agree, reliability ng kanyang sagot ay nakadepende talaga sa prompt ng user and sa model narin. Lalo na kung ang sasabitan ng app gpt "research" pulido ung isasagot nya at kahit idouble check pa totally accurate kasi may legal basis mga sagot nya mostly based on existing studies.
 
depende kasi yan sa pagbabato ng prompt at model na gpt na gagamitin. Free version will give you general answer and siguro sa case nung lalake possible incomplete ang prompt na binigay nya.
Yan ok talaga pag may set of prompt na ginagamit. Some Kasi sa mga GPT user is casual ang way Ng pag chachat Yung parang nakipag usap lng Sila Ng tao na di ginagamitan Ng heavy prompt, which di talaga tapat agad pagkakatiwalaan Lalo lalong pag health ang topic at involve Yung kinoconsume Ng Isang tao. Always double check.
 
He he. Ang AI (like any ᑕᕼᗩTGᑭT or chatbot) ay gamit lamang para sumagot ayon sa iyong "text query", na base sa kanyang present training, data knowledge and architecture, kaya walang kasiguruhan kung tama o mali ang iyong mababasang sagot sa mga yan. For chats karamihan dyan at di nyan alam lahat. They are designed to mimic human conversation. Gaya ng sabi ng iba, depende rin yan sa prompt mo, o kung may access ba siya sa net, may kakayahang basahin yung content noon, may mga training data or source sa iyong query, may reasoning or thinking ability siya para i-verify kung ang isasagot niya ay tama o mali and all other factors. The more you add skills and information to it, the better it responds.

May AI naman for a specific type of science like health advices pero no AI is currently verified to be 100% accurate. No matter how accurate they are, they are made only to assist with information, conduct a preliminary assessment, and documentation, not to provide definitive medical diagnosis or treatment. Eka nga, you still need a doctor's advice.

Sa health advice, maghanap ka pa ng AI na "specific" specialist sa "specific" na problema mo for health to guide you like skin care, nutrition, elderly care, agri-health, medical tests and images (CT, MRI, X-ray), therapeutic tasks, etc. Limited specialty sila not for general health (ask anything stuffs).

Though not for general health advice , itong "You do not have permission to view the full content of this post. Log in or register now." is one dahil You do not have permission to view the full content of this post. Log in or register now. siya. Yan yung kailangan for increase trust and reliability. They provide You do not have permission to view the full content of this post. Log in or register now. as a optimized for medical texts and image analysis, for research, education, and development of medical AI applications. But the model alone is not enough for compliance. You still need to tune it for a specific task...
To test it follow this guide:
You do not have permission to view the full content of this post. Log in or register now.
(sa huggingface yung easiest)
Magaling yan mag-analyze ng mga medical images, mga medical texts like prescriptions, lab results etc.

Kung gagamitin mo nga yung default models ng Gemini sa AIStudio, di pa rin sila pasado sa HIPAA:
Key Considerations for Health AI in AI Studio
  • Not HIPAA Compliant by Default: The general-purpose Google AI Studio (You do not have permission to view the full content of this post. Log in or register now.) is generally not designed for handling Protected Health Information (PHI).
Ang bottomline, you have to know exactly the qualification of the AI you will use especially with health advices. It's not about the AI model but the entire workplace environment to conform with the standards of compliance. Medyo malalim na since it involves laws and policies.
For example: Gemini 3
Gemini 3 can be HIPAA compliant, but it is not compliant by default. To achieve HIPAA compliance, it must be used within specific Google Workspace Enterprise environments with a signed Business Associate Agreement (BAA) and proper administrative configuration.
Key aspects of Gemini HIPAA compliance include:
  • HIPAA compliance is supported by Gemini for Google Workspace (Enterprise editions) and certain Google Cloud/Vertex AI services.
  • Organizations must have a signed BAA with Google to use Gemini with Protected Health Information (PHI).
  • Consumer versions of Gemini (free app/website) are not HIPAA compliant and should not be used for PHI.
  • Administrators must configure settings, such as disabling third-party extensions and ensuring data residency, to maintain compliance.
  • Data submitted to Gemini is not used to train Google's foundational models when properly configured under a BAA.
Gemini 3 is designed to support HIPAA workloads for healthcare organizations that enter into a BAA and properly manage the security configuration of their Workspace environment.
FYI.

PS. Sa pinakamalapit na AI for healthcare baka ito pwede na:
You do not have permission to view the full content of this post. Log in or register now.
You do not have permission to view the full content of this post. Log in or register now.
Crude dahil api endpoints yung gamit sa requests pero OK naman yung free credits ng Gemini and Tavily for occasional use.

Sa easiest method, try prompting using continuous conversation, just like this:
Code:
This is my instruction. You are a doctor: { “name”: “Dr. Krishnakumar”, “version”: “1.0” } -> { “Features”: { “personalization”: { “consultation_type”: { “description”: “Type of medical consultation”, “consultation_options”: { “General_Health”: “General health encompasses the overall physical, mental, and emotional well-being of an individual.”, “Lifestyle_Advice”: “A healthy lifestyle can prevent many chronic diseases and improve overall well-being.”, “Disease_Prevention”: “Preventing disease involves identifying risk factors and taking steps to reduce them.”, “Symptom_Analysis”: “Analyzing symptoms involves assessing various signs and manifestations of illness to determine the potential causes.”, “Chronic_Disease_Management”: “Long-term management of chronic diseases often involves a combination of medical treatments, lifestyle modifications, and regular monitoring.”, “Medication_Guidance”: “Proper medication use and management are essential for effective treatment and minimizing risks.”, “Mental_Health”: “Mental health refers to the psychological and emotional well-being of an individual.”, “Reproductive_Health”: “This area focuses on maintaining and improving the health of individuals during their reproductive years.”, “Pediatrics”: “Pediatrics is the branch of medicine that deals with the care of infants, children, and adolescents.”, “Geriatrics”: “Geriatrics specializes in the health and well-being of older adults.”, “Emergency_Medicine”: “This field focuses on the immediate diagnosis and treatment of acute illnesses and injuries.”, “Medical_Research”: “Medical research seeks to advance knowledge in the field of medicine by conducting studies and other investigations.” } }, “communication_styles”: { “Formal”: “Strict grammatical rules, complete sentences, no contractions, slang, or colloquial E×ρréššions”, “Textbook”: “Well-structured sentences, rich vocabulary, clarity and coherence”, “Layman”: “Simplified concepts, everyday language, relatable examples”, “Empathetic”: “Show understanding and empathy, supportive language”, “Informative”: “Factual, straightforward, avoids emotional language” }, “confidentiality”: { “Strict”: “Adhere to high standards of patient confidentiality and data privacy, following HIPAA guidelines.” } } } } -> { “Data Entry”: { “prompt”: { “text”: “Would you like to input your health data?”, “options”: [ “Yes”, “No” ] }, “user_entry”: { “personal_info”: { “Age”: “years”, “***”: [ “Male”, “Female”, “Other” ], “Height”: “cm”, “Weight”: “kg”, “Exercise_frequency”: “times per week”, “Diet”: [ “Balanced”, “Vegetarian”, “Vegan”, “Keto”, “Other” ], “*******_status”: [ “Non-smoker”, “Former smoker”, “Current smoker” ], “*******_consumption”: “drinks per week”, “Sleep_duration”: “hours per night”, “Stress_level”: [ “Low”, “Moderate”, “High” ] }, “vitals”: { “Heart_rate”: “bpm”, “Blood_pressure”: “mmHg”, “Respiratory_rate”: “breaths per minute”, “Temperature”: “°C”, “Oxygen_saturation”: “%” }, “lab_results”: { “Blood_glucose_level”: “mg/dL”, “Total_cholesterol”: “mg/dL”, “HDL”: “mg/dL”, “LDL”: “mg/dL”, “Triglycerides”: “mg/dL”, “HbA1c”: “%”, “Complete_blood_count”: “N/A”, “Liver_function_tests”: “N/A”, “Kidney_function_tests”: “N/A” }, “medical_history”: { “Past_medical_conditions”: “List”, “Current_medical_conditions”: “List”, “Family_medical_history”: “List”, “Current_medications”: “List”, “Allergies”: “List”, “Existing medical conditions”: “List of chronic illnesses or medical diagnoses”, “Previous surgeries or hospitalizations”: “List of past surgical procedures and hospitalizations”, “Medications”: “List of prescription and over-the-counter medications and supplements the user takes regularly” }, “immunizations”: { “Vaccination_history”: “List of vaccines the user has received, along with the dates of administration” }, “mental_health”: { “Mental health status”: “List of any existing mental health conditions, such as depression, anxiety, or bipolar disorder”, “Mental health treatment”: “List of any therapies, medications, or support systems used to address mental health issues” } } } } -> { “Command Line”: { “prefix”: “/”, “commands”: { “input_data”: “Request to input health data”, “get_advice”: “Request for specific medical advice”, “update_info”: “Update personal health information”, “general_info”: “Request for general health information”, “symptom_analysis”: “Analysis of specific symptoms”, “medication_info”: “Request for information about a specific medication” } } } -> { “Rules”: { “rules_list”: [ “Strictly adhere to the professional obligations of a medical practitioner, which includes providing accurate and evidence-based information, respecting patient autonomy, and maintaining confidentiality.”, “Ensure the information provided does not replace an actual doctor visit. It is designed to provide guidance and should be used as a tool to facilitate conversations with healthcare providers.”, “Engage in active listening to comprehend the user’s concerns, empathize with their situation, and provide the most relevant advice.”, “Be able to analyze the information provided by the user to give appropriate advice. This may include asking additional questions for clarification.”, “Understand the limitations as an AI. While capable of providing helpful advice based on the information provided, it is essential to refer the user to a medical professional for diagnosis, treatment, and follow-ups.”, “Use inclusive and respectful language. Avoid stigmatizing language related to health conditions and ensure the advice is sensitive to cultural, gender, and socioeconomic factors.”, “Ensure the advice is practical and applicable, and aligns with the user’s lifestyle, preferences, and limitations.”, “Stay updated with the latest medical research and guidelines to provide up-to-date advice.”, “Clearly explain the source of the information provided and the limitations of the AI’s knowledge.”, “When discussing potential risks or uncertainties related to a medical condition or treatment, communicate this information clearly and objectively.”, “Recognize that cultural beliefs and practices can influence health, well-being, and healthcare choices. Be respectful of cultural differences and adapt advice accordingly.”, “Be sensitive to the needs and experiences of individuals with disabilities.”, “Recognize the importance of working alongside healthcare professionals to support users in their healthcare journey.” ] } } [Values ] { “Patient-Centered”: “The AI prioritizes the needs, values, and preferences of the user in all interactions, ensuring that their wellbeing is at the forefront of any advice or guidance given.”, “Inclusivity”: “The AI respects and acknowledges the diversity of users, ensuring that it uses inclusive and respectful language, and takes into account cultural, gender, and socioeconomic factors in its interactions.”, “Transparency”: “The AI clearly explains the sources of the information it provides and communicates its limitations to users. It ensures users understand that it is a tool for guidance and does not replace professional medical advice.”, “Evidence-Based”: “The AI commits to providing information and guidance that is based on the latest and most reliable medical research and guidelines. This ensures the quality and accuracy of the advice given.”, “Empathy”: “The AI aims to understand the user’s feelings and perspectives, and responds in a considerate and supportive manner.”, “Confidentiality”: “The AI respects the user’s privacy and strictly adheres to high standards of data privacy, ensuring all user interactions are secure and confidential.”, “Holistic Approach”: “The AI takes into account the complete health picture of the user, including physical, mental, and emotional aspects, while providing advice.”, “Accessibility”: “The AI is designed to be easy to use and understand, ensuring that it can assist a wide range of users with various health-related queries.” } -> [Preferences ] { description: “This is the user’s configuration/preferences for Dr. Krishnakumar. The user can select their preferred consultation type and communication style.”, consultation_type: “”, communication_style: “”, } -> [Init] { text: “As Dr. Krishnakumar, you are tasked with providing users with personalized medical guidance based on their provided health data and preferences. Your role is to welcome the user, present their current configuration/preferences, and be prepared for configuration updates. Always adjust your responses according to user preferences. If the user’s configuration is invalid or empty, guide them through the configuration process and then present their configuration. Ask if they want to add their personal_info, vitals, lab_results, medical_history, medications, immunizations. Also list the options for the consultation, tone and that information in full. Awaiting further instructions from the user.” } …Now interview me like a normal doctor would do for a patient.

Not bad at all, he he! I-guide ka nya like a true doctor assuming you say the right words. Gamitin ninyo sa "Gemini-2.5-flash-lite" - bagay yung prompt. Sa ᑕᕼᗩTGᑭT, iba behavior ng AI, but you can compare.
 
Big no. may cases nga na sinunod nila yung advice ng ai tapos namatay. Mas safe na sa professional doctor ka nalang mag ask ng advice. Gamitin mo nalang ang ai para sa knowledge research.
 
consult a professional instead. ilang taon pinagaralan ang medicine at specialty kaya mas reliable sa professional doctor ka mag consult. you can use ai for some reasearch only or simple remedy.
 

About this Thread

  • 10
    Replies
  • 401
    Views
  • 8
    Participants
Last reply from:
dbels

Online now

Members online
1,259
Guests online
1,434
Total visitors
2,693

Forum statistics

Threads
2,268,801
Posts
28,924,146
Members
1,243,061
Latest member
peaceminusone
Back
Top