More

    Will Generative AI Help Us Solve The Climate Crisis (Or Will It Make It Worse)?

    Published on:

    You may be surprised to learn that AI has already proven to be a powerful weapon in the fight against climate change. For example, the United Nations is incorporating AI into its operations. UNOSAT Satellite Imagery Center – Use algorithms to analyze images of flooded areas and assess where disaster response teams need to go.

    AI can also help prevent future disasters and help vulnerable communities predict what will happen next. Indeed, the United Nations is working with vulnerable communities in Burundi, Chad and Sudan. Analyzing past environmental changes using AI Investigate mobile hotspots and build future predictions.

    However, since AI consumes a large amount of energy, to add Is it the same with climate issues? Yes, that's true too. (I never said there was an easy answer to the question in the title!) One of the biggest concerns is that with the rise of generative AI tools like ChatGPT, the energy costs of AI will only increase .

    This article considers the estimated energy costs of generative AI. Plus, we'll look at how new technologies can actually help solve or mitigate the climate crisis.

    ChatGPT (and friends) are energy devourers

    Back in 2019, researchers at the University of Massachusetts Amherst decided to investigate the carbon footprint of large-scale AI models of the kind that power today's tools like ChatGPT. They discovered that training a single AI model can: 5 times the lifetime emissions of the average American car (Including automobile manufacturing).

    That in itself is not ideal. It's not a dire number, but it's not ideal. But things have come a long way since then. At the time of our research, OpenAI's GPT large-scale language model was GPT-2. It has now reached GPT-4. This means that large-scale language models have grown significantly and become even bigger and more sophisticated. GPT-2 was trained with 1.5 billion parameters (the number of variables input into the model), while GPT-4 There are reportedly a whopping 1.8 trillion parameters. That's a big leap. Naturally, building such a very large model consumes more energy than the 2019 study found.

    And remember, that's just the energy needed to train the AI. There is also an energy associated with the people who actually use these tools. Asking a question to ChatGPT uses more energy than a standard internet search. This is a huge number compared to the roughly 180 million users ChatGPT has accumulated to date. The energy cost of these conversations is difficult to calculate, but some estimates suggest that ChatGPT searches may use power such as: 33,000 American households every day. every day. That's an amazing estimate. Also consumes each conversation with ChatGPT Approximately 50cl of water (small bottle).

    If tools like ChatGPT, Google Bard, and Microsoft's Bing AI chatbot continue to gain users, future numbers will no doubt become even more alarming.

    So does that mean generative AI will make the climate crisis worse?

    Well, it's not that simple. Because while generative AI tools are becoming a major source of carbon emissions, they can also help find the solutions needed to solve our biggest climate problems.

    There are multiple ways in which generative AI can help tackle climate change. For example, it can be used to better predict and manage energy demand. It can be used to generate more accurate forecasts of future energy production, allowing providers to allocate resources more effectively. This can be used to design better renewable systems – systems that are more efficient and scalable. It can also be used to design better carbon capture and storage solutions that capture and store emissions from factories and other industrial processes underground rather than releasing them into the atmosphere.

    Of course, this does not offset the huge energy costs associated with generative AI (and AI in general). It is imperative that AI companies commit to reducing their carbon footprint and invest in sustainable AI. But we cannot consider the climate costs of AI without recognizing its potential benefits.

    The bottom line is that the climate crisis is one of the most pressing problems humanity has ever faced. And generative AI is one of the most powerful technologies we have ever had access to. It is hoped that bringing the two together will lead to innovative solutions that will ultimately help fight climate change.

    Related

    Leave a Reply

    Please enter your comment!
    Please enter your name here