If you have specific health concerns or questions, it’s always best to consult a qualified healthcare professional instead of voice assistants for medical advice.
Voice assistants like Siri, Google Assistant, or Amazon Alexa can provide general information on medical topics and answer basic health questions. However, their accuracy in providing medical information may vary, and it’s important to understand their limitations.
The main issue is voice assistants deliver automatic responses and rely on online databases to provide information. So, they can offer general guidelines and information based on common medical knowledge. But, they’re not a substitute for professional medical advice.
Additionally, voice assistants are not capable of diagnosing medical conditions or providing personalized treatment recommendations.
It’s important to remember that medical conditions and treatments can vary widely based on individual circumstances, and relying solely on a voice assistant for medical advice may lead to incorrect or incomplete information. You need accurate and personalized guidance based on your unique situation.
Let’s examine a study that references our article comparing voice assistants within the context of voice recognition technology in 2023.
Evaluation of Replies to Voice Queries in Gynecologic Oncology
MDPI (Multidisciplinary Digital Publishing Institute) is a publisher of open access scientific journals. One of their recent publications really demonstrates what we’re talking about here.
The study began by noting women that receive news about a malignancy of gynecologic origin often have questions about their diagnosis. Totally natural. And these days, our first instinct is to pose these questions as voice queries to virtual assistants.
So, twenty-one evaluators analyzed voice assistant responses to select voice queries related to gynecologic oncology.
Here’s what they found.
For general queries that were not related to gynecologic oncology, Google provided the most correct audible replies (83.3% correct), followed by Alexa (66.7% correct), Siri (45.8% correct), and Cortana (20.8% correct).
Then, for gynecologic oncology-related queries, the accuracy was considerably lower: Google provided the most correct audible replies (18.1%), followed by Alexa (6.5%), Siri (5.5%), and Cortana (2.3%).
Therefore, “the accuracy of VA responses to voice queries related to gynecologic oncology was very low and inferior to queries that were of a general nature. Since treatment options for gynecologic malignancies are highly dependent on disease classifications such as stage, inaccurate informative responses by VAs could lead to misunderstandings by patients that cause them to refuse or withdraw from treatments that otherwise might have been beneficial.”
In summary, audible replies by voice assistants to voice queries related to gynecologic oncology have considerable room for improved accuracy. Overall, they recommend caution when using voice assistants to obtain information in gynecologic oncology.
But why is that? Because the data voice assistants provide is only as good as the data they’re fed.
Voice assistants for medical advice – What’s the source?
Voice assistants like Siri, Google Assistant, and Amazon Alexa obtain their answers from a combination of different sources and technologies.
Here are some of the key components that contribute to their knowledge and responses:
Web Search
Voice assistants often rely on web search engines like Google or Bing to retrieve information.
When you ask a question, the voice assistant may perform a web search to find relevant websites, articles, or other online sources that contain the information you’re looking for.
Knowledge Databases
Companies behind voice assistants maintain vast databases of information. These databases may include structured data – facts, figures, and general knowledge.
Voice assistants can retrieve answers directly from these databases when appropriate.
But they may not have the most up to date information.
APIs and Services
Voice assistants integrate with various application programming interfaces (APIs) and services to access specific information.
For example, they may connect to weather services, news platforms, or online encyclopedias to fetch real-time data or detailed information on specific topics.
User Data and Preferences
Voice assistants can learn from user interactions and personalize their responses based on individual preferences and past behavior.
They may store information about your preferences. That includes preferred music genres, commute routes, or frequently visited locations, to provide tailored recommendations and answers.
Device and Local Data
Voice assistants can access information stored locally on the device or within the ecosystem they operate.
This may include personal calendars, contacts, reminders, and other device-specific data that can be utilized to provide relevant responses.
Third-party Skills and Actions
Voice assistants often have a platform that allows third-party developers to create skills, actions, or apps.
These extensions provide additional functionality and access to specific services or databases beyond the built-in capabilities of the voice assistant.
For example, ordering a pizza or booking a ride through a voice assistant involves integration with third-party services.
None of the above, then, is conducive to getting medical information relevant to your specific case and context.
Further Limitations of Voice Assistants
As the study notes, the use of voice assistants internationally is a function of language and dialect.
As of March 2023, Google Assistant is available in 30 languages in 80 countries. Meanwhile, Siri supports 21 languages in 36 countries and Alexa is available in only 9 languages. That right there reduces accessibility and isn’t even accounting for different dialects among those specific languages.
More work is being done to improve these capabilities, and that involves huge amounts of data and accompanying annotation projects.
The rise of AI is relevant here too. Again, as the study notes, “Microsoft will launch an AI-powered Bing search engine on the Edge browser to ‘deliver better search, more complete answers, a new chat experience and the ability to generate content’”.
Additionally, Google plans a rollout of its AI chatbot named BARD, which is intended to enhance Google Search.
However, in a preview demonstration, Google BARD and Microsoft’s AI-powered Bing generated false information. And, at present, neither Microsoft nor Google has associated these AI-technologies with receiving and responding to voice instructions. So, that doesn’t really help us here, either.
The study concludes, “The new AI-powered Bing search Engine and Google BARD will need to be examined specifically for their validity in answering healthcare questions. And since they’re based on deep learning, they will need to be evaluated for a learning style that improves performance and keeps up with changes relevant to gynecologic oncology.”
In the end, talk to your doctor if you have questions about a medical diagnosis, and rely on your voice assistant to get you to your local pizza place.
We Can Help Improve Your Voice Technology Innovation
Voice assistants are continuously improving and evolving, but aren’t ready to distribute medical advice.
The specific sources and technologies they use to gather information may vary, and different voice assistant providers may employ unique approaches to generate responses.
All this is to say, if you think you might be sick, or have general questions about your health, you’re much better off calling your health care provider or visiting a clinic than asking for a smart device — at least for now.
If you need help with your voice recognition innovation, contact us.
As innovators in the data collection space, Summa Linguae Technologies offers flexible, customizable data solutions at scale that also evolve with your needs.
To see how we can help you with your data collection or annotation project, learn more about our data solutions here or contact us right away.