A study titled data-driven predictions of the time remaining until critical global warming thresholds are reached (2023), predicted a 1.5°C threshold somewhere between 2033 and 2035. Interestingly, artificial neural networks (ANNs) taught on climate model output. The application of AI is getting momentum like never before in all sorts of studies and research. AI was mentioned 30,000 times in the earning calls held by the end of 2023 and significant surge in 2024 & 2025.
If we look back, in 2016 major headlines about AI was Facebook shut down the AI systems. When facebook was developing sophisticated chatbots that could communicate like humans, however chatbots were interacting with each other in unusual ways and sometimes in code. Not one time. It happened several times and the communication pattern was difficult to decode for the researchers and programmers. Even though they were programmed to speak in English, chatbots were communicating differently and apparently forced Facebook to terminate the project.
The flip side of the story in 2024 is, It is estimated that the global artificial intelligence market will grow from $150.2 billion in 2023 to $1345.2 billion in 2030. Apart from the economics of AI, government bodies, corporate and researchers are making use of AI to tackle the most pressing issue of humankind – climate change.
In India, recognising the benefits of AI, NITI Aayog has formulated the National Strategy for Artificial Intelligence. Under this mission, healthcare, agriculture, education, smart cities and infrastructure and smart mobility and transportation have been identified as focus areas for AI intervention. Moreover, it is estimated that India can gain additional $957 billion or 15 percent of current gross value using AI.
Not just government, organisations are also reaping the benefits of AI. As per the research by the Capgemini Research Institute titled how artificial intelligence can power your climate action strategy, organisations have been able to reduce greenhouse gas (GHG) emissions by 13% and improved power efficiency by 11% since 2017 using AI.
Even though AI possesses several advantages, people still have concerns about it. Why?
There are two reasons for it. One, we cannot fully control every aspect of AI’s behaviour and second, we can’t always believe that an AI system stays in its limit and functions according to the code it is provided with.
For an AI system to work, firstly it undergoes a training phase. During this phase, large datasets are fed. Then, the system processes and analyses the dataset. Later, the system makes a decision based on the data it processes and analyses. After the training phase, it faces a test phase. In this phase, many real life examples are given to see whether the system performs rightly. During this entire process – feeding data to the test phase, three main challenges are faced.
First, during the test phase, there is a limit to the examples that can be provided to the system. Thus, examples are needed to see how machines react. Therefore, sometimes machines can be misled.
The next challenge is that the AI system can be corrupted during the training process. AI is made by humans. If the human who made the AI is dangerous and biassed, automatically the AI’s output and outcomes will be dangerous and biassed.
Similarly, after processing and analysing the data, the system can arrive at decisions which lack rationality, emotion and intelligence. Eventually, an advanced AI can foresee and defend itself if it finds humans as a threat.
In addition, manufacturing high end computation processors requires precious metals such as selenium, platinum and gold. Mining for such metals puts pressure on the earth and degrades the environment.
Moreover, running data centres for AI using non renewable sources of energy emits large amounts of carbon dioxide. It is estimated that over 250,000 pounds of carbon dioxide will be emitted for a single training of AI.
Having AI and other newer technologies gives us more real time mitigation against climate change. In the near future, researchers are able to discover alternative materials and renewable ways to make AI systems sustainable.
But, the real question is – what if the AI system supersedes human intelligence? What if AI systems manipulate climate data and create commotion among people? What steps can be initiated to address job losses caused by AI adoption?
Without addressing the above questions and going forward with AI, calls for a larger catastrophe than climate change. A legal framework needs to be formulated to protect individuals from surveillance and data mismanagement.
—————————————————————————————————————————–
Sources
- https://builtin.com/artificial-intelligence/facebook-shuts-down-a
- https://www.accenture.com/content/dam/accenture/final/accenture-com/document-2/Accenture-Tech-Vision-2024.pdf
- https://www.pnas.org/doi/10.1073/pnas.2207183120
- https://niti.gov.in/sites/default/files/2019-01/NationalStrategy-for-AI-Discussion-Paper.pdf
- https://www.niti.gov.in/sites/default/files/2021-02/Responsible-AI-22022021.pdf
- https://www.sciencealert.com/researchers-say-itll-be-impossible-to-control-a-super-intelligent-ai#:~:text=The%20answer%3F,to%20create%20such%20a%20simulation.
- https://builtin.com/artificial-intelligence/facebook-shuts-down-ai
- https://www.weforum.org/agenda/2016/10/top-10-ethical-issues-in-artificial-intelligence/
- https://dataconomy.com/2022/04/risks-and-benefits-of-artificial-intelligence/#Misuse_leading_to_threats
- https://materion.com/resource-center/newsletters/coating-materials-news/minimizing-costs-of-precious-metals-in-compound-semiconductors#:~:text=Gold%20(Au)%20and%20Platinum%20
- https://www.cfr.org/blog/artificial-intelligences-environmental-costs-and-promise#:~:text=Training%20a%20single%20AI%20system,comparable%20to%20the%20aviation%20industry.
- https://time.com/6266923/ai-eliezer-yudkowsky-open-letter-not-enough/
Featured Photo: Photo by Google DeepMind via pexels
