Beyond ChatGPT: IBM’s Vision for the True Potential of AI

AI Beyond ChatGPT: IBM’s Tangible Impact on Climate and Beyond

Introduction

In the rapidly evolving landscape of artificial intelligence (AI), IBM stands out as a pioneer, not just in pushing the boundaries of technology but in leveraging AI to address pressing global issues. While there’s ongoing discourse on the risks associated with advanced AI models like ChatGPT, IBM takes a pragmatic approach by focusing on tangible benefits. In this article, we delve into IBM’s recent strides in harnessing geospatial AI to combat climate change and its commitment to making a positive impact on our planet.

AI’s Role in Climate Change Mitigation

As concerns about the environmental footprint of Large Language Models (LLMs) grow, IBM has taken a distinct path, recognising the limitations of models like ChatGPT in directly addressing climate-related challenges. Dr. Sasha Luccioni, a prominent figure in ethical AI, points out that LLMs may not be the panacea for reinventing material science and saving the climate. IBM, however, doesn’t dismiss the potential of AI; instead, it shifts the focus to smaller, more precise models with reduced carbon footprints.

IBM’s Geospatial AI Model: A Game-Changer

In August, IBM underscored its commitment to environmental stewardship by open-sourcing its geospatial AI model, developed in collaboration with NASA, on Hugging Face. This move marks a significant stride in democratising access to cutting-edge technology to benefit the planet.

Alessandro Curioni, IBM Fellow and Vice President of accelerated discovery, emphasises the transformative potential of AI foundation models utilising geospatial data. According to Curioni, these models can revolutionise our understanding, preparation, and response to climate-related events with unprecedented speed and accuracy.

Mapping Urban Heat Islands: A Collaborative Effort

One compelling application of IBM’s geospatial model involves a collaboration with the Mohamed Bin Zayed University of Artificial Intelligence. In this initiative, scientists leverage the power of IBM’s model to map urban heat islands, areas with significantly higher temperatures than their surroundings. Focusing initially on Abu Dhabi, the project has successfully reduced the heat island effect by more than 5.4 degrees Fahrenheit, showcasing the tangible impact of AI in real-world scenarios.

Reforestation Campaign Powered by AI

IBM’s commitment extends beyond mapping to on-the-ground action. Partnering with the Kenyan government, IBM supports a reforestation plan aiming to plant 15 billion trees across the country by 2032. The geospatial model plays a pivotal role in tracking and visualising tree planting and growth progress, enabling efficient mobilisation of efforts on a massive scale.

Automated Climate Risk Assessment with Royal HaskoningDHV

Collaborating with Royal HaskoningDHV, a leading engineering company, IBM introduces a groundbreaking service that utilises its geospatial model to automate and scale climate risk assessments. The initial focus is on analysing the short-term impact of extreme weather on aviation operations, as well as the long-term implications of climate change on future airport infrastructure. This marks a crucial step towards enhancing climate resilience in critical sectors.

Geospatial Mapping for Climate Resilience in the U.K.

IBM’s impact transcends borders, as it actively collaborates with various sectors in the United Kingdom to bolster the country’s climate resilience. One notable effort involves geospatial mapping of urban areas to identify optimal locations for tree planting, strategically reducing the risk of flooding. Chloe Treger, the lead of TreesAI U.K., attests to the effectiveness of IBM’s model in assessing scenarios and making a compelling business case for tree planting and maintenance.

The Power of Geospatial Models

IBM’s geospatial model, trained over a year on extensive U.S. data, proves to be a versatile tool in addressing climate-related challenges. Boasting 100 million parameters, a fraction of ChatGPT’s 175 billion, this model offers a cost-effective approach to tracking deforestation and greenhouse gas emissions. Konstantin Klemmer, a geospatial machine learning researcher, acknowledges the immense potential of such models in decoding patterns in the ever-expanding realm of geospatial data.

Despite potential training costs and CO2 emissions associated with large-scale geospatial models, Klemmer asserts that they present a more economical and faster alternative for making forecasts compared to traditional numerical simulators. This shift signifies a strategic tradeoff between accuracy and efficiency, with geospatial foundation models emerging as powerful tools in the battle against climate change.

Conclusion

In conclusion, IBM’s foray into geospatial AI represents a paradigm shift in addressing climate change. While the discourse on AI’s role in environmental sustainability continues, IBM’s tangible initiatives underscore the transformative power of targeted, ethical AI applications. From mapping urban heat islands to supporting reforestation campaigns, IBM’s geospatial model is not just a technological marvel; it’s a catalyst for meaningful change. As the world grapples with the urgent need for climate action, IBM stands at the forefront, showcasing the potential of geospatial AI to make a lasting impact.