In a revolutionary move, the World Health Organization (WHO) has unveiled groundbreaking ethical guidelines focusing on the use of large multi-modal models (LMMs) in healthcare. These guidelines are a compass for governments, technology companies, and healthcare providers, emphasizing the responsible and equitable utilization of AI technologies in the medical realm.
The World Health Organization (WHO) is releasing new guidance on the ethics and governance of large multi-modal models (LMMs) – a type of fast-growing generative artificial intelligence (AI) technology with applications across health care.
Riding the Wave: The Rise of LMMs
LMMs are not your run-of-the-mill AI technology; they’re the rockstars of the artificial intelligence world, capable of accepting a plethora of data inputs like text, videos, and images, and spitting out diverse outputs that defy the boundaries of their inputs. Think ChatGPT, Bard, and Bert – names that entered the public consciousness like a tidal wave in 2023.
“Generative AI technologies have the potential to improve healthcare, but only if those who develop, regulate, and use these technologies identify and fully account for the associated risks,” says Dr. Jeremy Farrar, WHO Chief Scientist. “We need transparent information and policies to manage the design, development, and use of LMMs to achieve better health outcomes and overcome persisting health inequities.”
Unveiling the Pandora’s Box: Potential Benefits and Risks
The new WHO guidance shines a light on five broad applications of LMMs in healthcare, opening avenues for:
- Diagnosis and Clinical Care: Responding to patients’ written queries.
- Patient-Guided Use: Investigating symptoms and treatment based on patient input.
- Clerical and Administrative Tasks: Documenting and summarizing patient visits within electronic health records.
- Medical and Nursing Education: Providing trainees with simulated patient encounters.
- Scientific Research and Drug Development: Identifying new compounds for medical advancements.
The Flip Side: Risks and Challenges
While LMMs promise a brave new world in healthcare, the guidance doesn’t shy away from addressing the elephant in the room – the risks. These include:
- False or Inaccurate Information: LMMs may produce information that is false, inaccurate, biased, or incomplete, posing a threat to those making health decisions based on such data.
- Quality and Bias Issues: LMMs can be trained on poor-quality or biased data, leading to skewed results based on factors like race, ethnicity, sex, gender identity, or age.
- Broader Health System Risks: Accessibility and affordability of the best-performing LMMs could create disparities in healthcare. The notorious ‘automation bias’ may creep in, where healthcare professionals and patients overlook errors or delegate critical choices to LMMs.
Navigating the Waters: Key Recommendations
Governments Take the Helm
The new WHO guidance serves as a compass for governments, urging them to:
- Invest in Infrastructure: Provide public infrastructure, including computing power and data sets, accessible to developers across sectors while adhering to ethical principles.
- Regulate with Purpose: Use laws, policies, and regulations to ensure that LMMs meet ethical obligations and human rights standards, regardless of associated risks or benefits.
- Establish Regulatory Oversight: Assign regulatory agencies to assess and approve LMMs for healthcare use, ensuring adherence to ethical and human rights standards.
- Post-Release Auditing: Introduce mandatory post-release auditing and impact assessments by independent third parties, covering data protection and human rights, with outcomes and impacts disaggregated by user type.
Developers, Strap In!
The guidance also throws the spotlight on LMM developers, emphasizing that:
- Diverse Inclusion: LMMs should not be designed in silos by scientists and engineers. Potential users, medical providers, researchers, healthcare professionals, and patients must be engaged from the early stages of development to voice concerns and provide input.
- Well-Defined Tasks: LMMs should be designed to perform well-defined tasks with accuracy and reliability, ensuring they enhance health systems and prioritize patient interests.
FAQs: Demystifying the WHO Guidelines
Q1: Why is WHO emphasizing LMMs now?
The WHO recognizes the explosive growth of LMMs in healthcare and aims to ensure their responsible use, addressing associated risks and promoting equitable access.
Q2: What are the primary risks associated with LMMs?
Risks include producing false or biased information, poor-quality training data, and broader health system challenges like accessibility and affordability disparities.
Q3: How can governments regulate LMMs effectively?
Governments should invest in infrastructure, regulate with ethical principles, establish regulatory oversight, and implement post-release auditing by independent third parties.
In Conclusion: Navigating the AI Seas Responsibly
As LMMs continue to reshape the landscape of healthcare, the WHO’s ethical guidelines stand as a lighthouse, guiding governments, developers, and stakeholders through the turbulent waters of AI. With transparency, inclusivity, and a commitment to ethical standards, we can harness the potential of LMMs to revolutionize healthcare without compromising on safety or equity. The WHO has sounded the alarm – now it’s up to us to navigate the AI seas responsibly and ensure a healthier future for all.