Artificial intelligence (AI) has exploded onto the global stage in recent years. While some view its rapid advancement with anxiety, AI holds immense potential to benefit humanity. In their book, “AI for Good” (Wiley, 2024), Juan M. Lavista Ferres and William B. Weeks, senior directors at Microsoft’s AI for Good Research Lab, highlight numerous projects worldwide showcasing AI’s positive impact.
They demonstrate how AI is contributing to sustainability projects like using satellites to monitor whales or mapping glacial lakes. It’s also proving valuable in disaster relief, as seen in the response to the 2023 earthquake in Turkey. Furthermore, AI is being used for social good, combatting the spread of misinformation online. The health benefits of AI are equally significant, including research into long-term COVID-19 effects, management of pancreatic cysts, and detection of leprosy in vulnerable populations.
The authors delve into the recent surge of large language models (LLMs) like ChatGPT and Claude 3, which have become prominent players in the AI landscape. They discuss the remarkable impact these systems are having on the world.
At the heart of today’s linguistic technologies, such as GPT, lies the concept of a language model. Imagine starting a sentence with “This morning I woke up and saw a beautiful blue ____.” What should follow? A language model predicts the continuation based on probabilities derived from vast amounts of textual data. Words like “sky” might be highly probable continuations. However, the model’s sophistication allows it to consider a diverse range of possibilities, such as “bird” or “car,” each with a specific probability, showcasing its nuanced understanding of different contexts.
Despite their limitations, the influence of LLMs has been remarkable, particularly in the latter part of 2023. GPT-4 has achieved significant milestones, including passing both the multiple-choice and written sections of the bar exam. These models’ true strength lies in their ability to learn from the vast source of information that is the World Wide Web. This immense resource contains a significant portion of our collective human knowledge and is arguably the most important dataset in the world. By training on this enormous dataset, LLMs can build a representation of the world that replicates the complex relationships found within human understanding.
While these advanced LLMs don’t possess real understanding or consciousness, their capacity to process and emulate human language and thought is immensely powerful. As we continue to evolve and refine these technologies, it’s crucial to understand their capabilities, limitations, and the ethical implications they bring.
Beyond healthcare diagnostics, where AI’s power is widely discussed, LLMs are proving invaluable as language aids for non-native speakers. This resonates deeply with the authors, one of whom is a non-native English speaker leading a research lab with a majority of members from the Global South, who are also non-native speakers. With approximately 95% of research published in English and only 4.7% of the global population being native English speakers, mastering English is crucial for many researchers. LLMs like GPT offer a path to confidently writing in English for everyone.
LLMs are not a panacea, but they have the potential to bridge language gaps in remarkable ways. A good translation tool doesn’t just provide a literal translation of words between two languages. It must also convey the meaning, tone, cultural connotation, and context. LLMs work by refining text that may not be well structured, converting it into expressions akin to those of a native speaker. This is especially vital for non-native speakers who often grapple with the nuances of English grammar and syntax. The model assists not just by ensuring grammatical accuracy, but also by enhancing vocabulary to match the quality of native English publications.
The authors also highlight the potential of LLMs to democratize coding. Learning to code is akin to mastering a new language, serving as our interface to programming computers. While there has been progress in increasing the number of people learning to code, it’s challenging to foresee a radical increase in the coming decades. LLMs, however, could bring about a substantial shift. Advanced systems like GPT-4 have the capability to translate natural language into actual programming languages. This empowers individuals to write programs and automate processes in their native tongues, be it English, Spanish, Mandarin, or others. This technology has the potential to democratize programming, extending its reach to hundreds of millions worldwide and bridging the gap between those who can code and those who cannot.
In the realm of medicine, LLMs are demonstrating remarkable potential. A study published in JAMA International Medicine in April 2023 compared the responses of physicians to those of GPT-4 when answering patient queries. The study found that GPT-4 not only provided more accurate answers than the doctors but also demonstrated greater empathy. While GPT-4 hasn’t been extensively trained on medical knowledge, particularly those behind paywalls, and wasn’t specifically trained for medical scenarios, its performance highlights its potential impact. With approximately 4 billion people lacking access to doctors, these AI models could significantly impact global healthcare disparities. While these models are not intended to replace doctors, they could provide accurate responses to human inquiries, allowing doctors to focus on areas where they excel. The promising results suggest a path forward in addressing global healthcare disparities.
The authors’ book, “AI for Good: Applications in Sustainability, Humanitarian Action, and Health,” offers a compelling and insightful discussion of how scientists are using AI to tackle some of the world’s most pressing challenges. It underscores the transformative potential of AI in improving lives and creating a better future for all.
