The World Health Organization (WHO) has ventured into the realm of artificial intelligence (AI) by utilizing a human-like avatar, SARAH, to provide basic health information to the public. Tailored to assist users around the clock in eight languages, SARAH is designed to offer guidance on various health topics such as mental health, tobacco cessation, and maintaining a balanced diet.
Understanding SARAH’s Capabilities and Limitations
The virtual health assistant is not without its limitations, occasionally echoing inaccurate information. WHO advises users that SARAH, sourced from data as recent as September 2021, may not always furnish accurate responses. This is due, in part, to training on outdated data, which can lead to erroneous replies akin to ‘hallucinations’ within the AI model.
SARAH’s operational domain is carefully set to steer clear of diagnosis, akin to platforms like WebMD and Google. If inquired about topics outside WHO’s scope, such as specific medications, SARAH redirects users to the WHO website or suggests consulting a health professional.
The Medical Community’s Reaction
Dr. Ramin Javan, a radiologist and researcher at George Washington University, recognizes the content’s limitations but appreciates the conservative and cautious approach at this nascent stage. WHO hopes to enhance SARAH’s performance by collaborating with researchers and government entities to ensure it can suggest healthier lifestyle choices backed by accurate public health statistics.
Despite the progress, Alain Labrique, director of WHO’s Digital Health and Innovation section, stresses that this technology is not yet ready to substitute professional consultations.
Data Privacy and Security Measures in Place
In its commitment to privacy, SARAH anonymizes each session and does not retain personal data such as IP addresses. Although WHO’s Jaime Guerra assures that anonymity reduces the threat of data breaches, concerns linger over the potential dangers of using open-source databases like GPT which could attract cybercriminals.
Potential for Future Development
SARAH, building upon WHO’s earlier virtual health assistant project ‘Florence’, provides basic information on topics such as COVID-19 and tobacco use. While Soul Machines, the company responsible for crafting the avatar, does not access SARAH data, it utilizes GPT to enhance outcomes and user experience.
In response to the pressing need for data transparency and safety, WHO has published ethical guidelines for health-related AI models for its governmental partners. While the avatar’s appearance and the gender presentation are adjustable, SARAH remains focused on its mission: to support healthy living without ascribing to a specific gender identity.
Relevance of Artificial Intelligence in Health Communications
The introduction of AI avatars like SARAH in healthcare is a testament to the rapid digitization of health information dissemination. SARAH’s development represents a growing trend where health organizations leverage technology to increase accessibility to health education and support. The use of AI in this context is key to providing scalable solutions to disseminate information amidst global challenges such as the COVID-19 pandemic.
Important Questions and Answers
1. Q: What is the purpose of SARAH?
A: SARAH is designed to offer basic health guidance and information to the public on various health topics.
2. Q: Does SARAH provide medical diagnoses?
A: No, SARAH is programmed to avoid diagnosis and instead redirects users to professional consultations where necessary.
3. Q: How does WHO ensure the privacy and security of users interacting with SARAH?
A: WHO states that SARAH anonymizes each session and does not store personal identifying information, to mitigate data privacy and security risks.
Key Challenges and Controversies
– Data Accuracy: One of the primary challenges is ensuring that the information provided by SARAH is up-to-date and accurate. Since SARAH’s data is only as recent as September 2021, there is a risk for outdated or incorrect guidance.
– Privacy Concerns: Although SARAH anonymizes data, there are still general concerns about user privacy and the security of the data in AI tools that utilize open-source databases.
– Depersonalization of Care: Another critique is that digital tools may lead to a depersonalization of care, potentially undermining the patient-clinician relationship if users start to rely excessively on AI for health guidance.
Advantages and Disadvantages of AI in Health Guidance
Advantages:
– Accessibility: Provides 24/7 health information access in multiple languages.
– Scalability: Capable of reaching a wider audience compared to one-on-one health consultations.
– Anonymity: Can offer a level of comfort for users who wish to inquire about sensitive health issues privately.
Disadvantages:
– Accuracy: AI systems can propagate inaccuracies if trained on outdated or incorrect data.
– Limitation of Scope: AI tools like SARAH are not substitutes for professional medical diagnosis or treatment.
– Cybersecurity Risks: Platforms using open-source data might be targets for cyberattacks, leading to broader data privacy concerns.
For more information about the World Health Organization and its initiatives, you can visit the main WHO website at World Health Organization.