5 selected ways to utilize AI for achieving SDGs. What is sustainable green AI.

1. Efforts to Address Environmental Issues Using AI

 

Introducing examples of utilizing AI to address environmental issues.

1-1. Case 1: Monitoring Deforestation with Sound Detection

Rainforest Connection, a non-profit organization based in the United States, provides a system for monitoring illegal deforestation by detecting sound. By installing smartphones as microphones in the forest and using AI on the cloud to identify the sounds of chainsaws and trucks used for transporting wood, this unique system allows for faster response times than traditional monitoring systems by contacting local conservation officers through the use of the local mobile phone network.

>>Old smartphones play an active role, listening to the "sound" of illegal logging

1-2. Case 2: Understanding the Ecology of Sperm Whales

AI technology is being used to protect the endangered species of the humpback whale. With the AI developed by NOAA (National Oceanic and Atmospheric Administration) and Google, it is possible to extract specific sounds from environmental noise. This technology allows for the recording of humpback whale sounds in the ocean, even in environments with various types of noise. By accurately understanding the timing and location of humpback whale activity, it is contributing to the advancement of conservation efforts.

Google's efforts to listen to the "voices" of whales using AI for understanding the ecology of sperm whales is currently underway - GIGAZINE

1-3. Case 3: Reducing Agricultural Water Consumption with AI Weather Forecast

The Yield, an Australian startup company, offers a highly accurate weather forecasting service using AI technology. This service has successfully streamlined the schedules of farmers and reduced the consumption of agricultural water. It has also led to an increase in harvest yields.
The company is also working on the development of smart agriculture using AI in Japan through a partnership with Yamaha Motor.

>>Yamaha Motor and Smart Agriculture Sign Joint Development Agreement for Bold Startup

1-4. Case 4: Streamlining Fishing Vessel Operation Schedules to Reduce Fuel Consumption

Our company also offers services specialized in increasing production for oyster farming. By preventing cancellations and changes in plans due to weather conditions, we not only improve production, but also contribute to reducing fuel consumption used for fishing boat operations by streamlining schedules.

>>Internet of Oysters: The Yield delivers sunnier results for Australian oyster farmers (English)

In Japan, UT Group Co., Ltd. and Sharp Corporation are conducting a demonstration experiment of "Smart Oyster Farming" using AI at the University of Tokyo.

Started demonstration experiment of "Smart Oyster Farming" utilizing AI/IoT

1-5. Case 5: Monitoring Earth's Resources with Satellite Images and AI

The introduction of satellite images and AI is effective for monitoring Earth's resources. It is extremely difficult to monitor vast areas only with human visual inspection. By monitoring deforestation, natural gas emissions, and melting of permafrost in the Arctic region based on images from above, we are promoting effective environmental protection activities.

>>From Earth to AI: 3 Startups Monitor the Environment Using Deep Learning

1-6. Case 6: Automatic Identification of Waste in Incinerators using AI

In "waste incineration power generation," which generates electricity from the heat generated by burning waste, the waste used as raw material is originally wood or food that has absorbed carbon dioxide from the atmosphere. Although the carbon dioxide generated during incineration returns to the atmosphere, it is a carbon-neutral power generation method in terms of total amount. The challenge of waste incineration power generation is the decrease in energy recovery rate due to unstable incineration caused by the uneven mixing of waste used as fuel. In order to solve this challenge, it was necessary to mix the waste uniformly, but in the past, it relied on the experience of workers. By introducing AI that can identify waste into the device for mixing, more efficient mixing can be achieved. Furthermore, by introducing AI for control during combustion, the energy recovery rate can be further increased.

Improve the efficiency of waste incineration power generation with AI

1-7. Case 7: AI Traffic Congestion Prediction Model

It has been found that CO2 emissions during traffic congestion are about twice as high as during normal times. By using a traffic prediction model, it is possible to reduce CO2 emissions by preventing traffic congestion in advance. Various companies are developing prediction models, and for example, Geo Technologies Co., Ltd. has successfully developed a model that can predict traffic congestion in 5-minute intervals even on general roads. If utilized, it is also expected to reduce CO2 emissions by avoiding traffic congestion.

Successfully developed "AI Traffic Prediction Model" that can predict in 5-minute increments even on general roads.

2. The Problem of Energy Consumption Caused by AI Development

 

This section will focus on the negative aspects of AI development on the environment.

2-1. Relationship between AI Development and Environmental Impact

The background of AI development involves processes and methods such as machine learning and deep learning. The issue of environmental impact caused by AI development is due to these processes. In order to achieve more accurate AI, a large amount of data must be processed and machine learning must be performed. The use of a large number of processors requires an increase in power consumption and the use of power for cooling. If companies that have not previously implemented AI introduce it, their power consumption will increase. As a result, the development and implementation of AI will lead to an increase in CO2 emissions.

CO2 emissions from AI development are five times that of automobiles.

In 2019, a life cycle assessment (LCA) was conducted for an AI development at the University of Massachusetts in the United States. This deep learning model for natural language processing, called Transformer, uses over 200 million parameters for training. The CO2 emissions generated during this process are equivalent to about 5 times the amount emitted by a car from production to disposal, and 300 times more than a round-trip flight between New York and San Francisco. The increase in energy consumption for AI development has become a significant issue that cannot be ignored.

Source: MIT Technology Review 2019 "Training a single AI model can emit as much carbon as five cars in their lifetimes" Created by the author.

2-3. AI Generation and CO2 Emissions

The use of generative AI, represented by ChatGPT, is rapidly spreading. The development of generative AI requires large-scale servers, and a report from Stanford University in April 2023 revealed that data centers that house these servers consume a large amount of electricity.
According to the report, GPT-3 emits 502 tons of CO₂ per year, which is about 507 times more than the CO₂ emissions per passenger when flying round trip between New York and San Francisco (0.99 tons per person).
It has also been found that a large amount of water is consumed for server cooling in the development process. It is estimated that asking GPT-3 25 to 50 questions consumes about one bottle of water per user. As AI advances and becomes more widespread, there is a disadvantage that the burden on the environment will also increase proportionally.

>>Chat GPT Accelerates Global Warming

3. Sustainable AI Development with Green AI

3-1. What is Green AI?

The development of AI is driven by the increasing environmental burden. On the other hand, it is also a fact that AI is having a positive impact on addressing environmental issues. One solution that is gaining attention to improve this conflicting situation is Green AI.

 

Green AI refers to AI that has been developed through research, methods, or approaches that aim to operate AI with the least amount of energy possible. This means taking a comprehensive approach towards sustainable AI development, rather than compromising on the quality or accuracy of AI.
Examples of such initiatives include:

 

・Choose a location that is easy to receive renewable energy when setting up a data center.
・When using public cloud, choose a provider that actively utilizes renewable energy.
・In the processing and development of AI, pursue more efficient algorithms to reduce processor load.

3-2. What is Red AI?

While Green AI focuses on comprehensive perspectives, AI development that only pursues accuracy without such perspectives is called Red AI. It is a method that pours maximum energy to push the limits of computer processing power. From the perspective of SDGs, a review before operation is required.

3-3. Contribute to carbon neutrality simply by using the cloud

Public cloud data centers have already announced their operational policies of using renewable energy for power supply and reducing waste. Choosing these data centers will directly contribute to achieving SDGs.

 

Microsoft has announced that by using their Azure data centers, companies can reduce their carbon emissions by up to 98% compared to on-premises operations.

>>Microsoft to provide data infrastructure related to CO2 emissions in the first half of the year

 

Even in domestic companies' data centers, each company has set specific goals towards achieving carbon neutrality.

>>No Time to Wait: Decarbonizing Data Centers

 

In the future of AI development, consideration for SDGs in the development process is required, not just the functionality and accuracy of the released product. In order to balance the development of AI and the achievement of SDGs, let's start by reviewing work efficiency and selecting data centers.

4. For inquiries about utilizing AI, please contact Human Science Co., Ltd.

4-1. Rich track record of creating 48 million pieces of teacher data

At Human Science, we are involved in AI model development projects in various industries such as natural language processing, medical support, automotive, IT, manufacturing, and construction. Through direct transactions with numerous companies including GAFAM, we have provided over 48 million high-quality training data. We handle various annotation projects regardless of industry, from small-scale projects to large-scale projects with 150 annotators. If your company is interested in implementing AI models but unsure of where to start, please consult with us.

>>Data Annotation Service by Human Science Co., Ltd.

4-2.Resource Management without Using Crowdsourcing

At Human Science, we do not use crowdsourcing and instead directly contract with workers to manage projects. We carefully assess each member's practical experience and evaluations from previous projects to form a team that can perform to the best of their abilities.

4-3. Utilizing the Latest Data Annotation Tools

One of the annotation tools introduced by Human Science, AnnoFab, allows customers to check progress and provide feedback on the cloud even during project execution. By not allowing work data to be saved on local machines, we also consider security.

4-4. Equipped with a security room within the company

At Human Science, we have a security room that meets the ISMS standards in our Shinjuku office. We can handle highly confidential projects on-site. We consider ensuring confidentiality to be extremely important for all projects. We continuously provide security education to our staff and pay close attention to handling information and data, even for remote projects.