The image above is on The Generative AI Race Has a Dirty Secret, credit to WIRED
AI chatbots and image generators run on thousands of computers housed in data centers like this Google facility in Oregon. Tony Webster/Wikimedia, CC BY-SA
Generative AI is the hot new technology behind chatbots and image generators. But how hot is it making the planet?
As an AI researcher, I often worry about the energy costs of building artificial intelligence models. The more powerful the AI, the more energy it takes. What does the emergence of increasingly more powerful generative AI models mean for society’s future carbon footprint?
“Generative” refers to the ability of an AI algorithm to produce complex data. The alternative is “discriminative” AI, which chooses between a fixed number of options and produces just a single number. An example of a discriminative output is choosing whether to approve a loan application.
Generative AI can create much more complex outputs, such as a sentence, a paragraph, an image or even a short video. It has long been used in applications like smart speakers to generate audio responses, or in autocomplete to suggest a search query. However, it only recently gained the ability to generate humanlike language and realistic photos.
Using more power than ever
The exact energy cost of a single AI model is difficult to estimate, and includes the energy used to manufacture the computing equipment, create the model and use the model in production. In 2019, researchers found that creating a generative AI model called BERT with 110 million parameters consumed the energy of a round-trip transcontinental flight for one person. The number of parameters refers to the size of the model, with larger models generally being more skilled. Researchers estimated that creating the much larger GPT-3, which has 175 billion parameters, consumed 1,287 megawatt hours of electricity and generated 552 tons of carbon dioxide equivalent, the equivalent of 123 gasoline-powered passenger vehicles driven for one year. And that’s just for getting the model ready to launch, before any consumers start using it.
Size is not the only predictor of carbon emissions. The open-access BLOOM model, developed by the BigScience project in France, is similar in size to GPT-3 but has a much lower carbon footprint, consuming 433 MWh of electricity in generating 30 tons of CO2eq. A study by Google found that for the same size, using a more efficient model architecture and processor and a greener data center can reduce the carbon footprint by 100 to 1,000 times.
Larger models do use more energy during their deployment. There is limited data on the carbon footprint of a single generative AI query, but some industry figures estimate it to be four to five times higher than that of a search engine query. As chatbots and image generators become more popular, and as Google and Microsoft incorporate AI language models into their search engines, the number of queries they receive each day could grow exponentially.
AI chatbots, search engines and image generators are rapidly going mainstream, adding to AI’s carbon footprint. AP Photo/Steve Helber
AI bots for search
A few years ago, not many people outside of research labs were using models like BERT or GPT. That changed on Nov. 30, 2022, when OpenAI released ChatGPT. According to the latest available data, ChatGPT had over 1.5 billion visits in March 2023. Microsoft incorporated ChatGPT into its search engine, Bing, and made it available to everyone on May 4, 2023. If chatbots become as popular as search engines, the energy costs of deploying the AIs could really add up. But AI assistants have many more uses than just search, such as writing documents, solving math problems and creating marketing campaigns.
Another problem is that AI models need to be continually updated. For example, ChatGPT was only trained on data from up to 2021, so it does not know about anything that happened since then. The carbon footprint of creating ChatGPT isn’t public information, but it is likely much higher than that of GPT-3. If it had to be recreated on a regular basis to update its knowledge, the energy costs would grow even larger.
One upside is that asking a chatbot can be a more direct way to get information than using a search engine. Instead of getting a page full of links, you get a direct answer as you would from a human, assuming issues of accuracy are mitigated. Getting to the information quicker could potentially offset the increased energy use compared to a search engine.
Ways forward
The future is hard to predict, but large generative AI models are here to stay, and people will probably increasingly turn to them for information. For example, if a student needs help solving a math problem now, they ask a tutor or a friend, or consult a textbook. In the future, they will probably ask a chatbot. The same goes for other expert knowledge such as legal advice or medical expertise.
While a single large AI model is not going to ruin the environment, if a thousand companies develop slightly different AI bots for different purposes, each used by millions of customers, the energy use could become an issue. More research is needed to make generative AI more efficient. The good news is that AI can run on renewable energy. By bringing the computation to where green energy is more abundant, or scheduling computation for times of day when renewable energy is more available, emissions can be reduced by a factor of 30 to 40, compared to using a grid dominated by fossil fuels.
Finally, societal pressure may be helpful to encourage companies and research labs to publish the carbon footprints of their AI models, as some already do. In the future, perhaps consumers could even use this information to choose a “greener” chatbot.
A smart city uses digitalisation-supported information and communication technology (ICT) in its diverse operational exercises, shares information and provides better governance.: Constructing materially smarter cities on Elkem.
In 2050 close to 70 percent of the world’s population is expected to live in cities and the need for efficient infrastructure will increase. Did you know that the materials used on satellites and space applications play a crucial role in enabling smart and safe cities of the future?
There are different definitions of what a smart city actually is. As a general interpretation, however, consensus seems to align around that the term says something about the degree to which traditional networks and services are made more efficient with use of digital and telecommunication technologies – for the benefit of its inhabitants and businesses
Smart cities use Internet of Things (IoT) devices, like sensors, lights, and meters to collect and analyse data. The cities can then use this data to improve infrastructure, public utilities and services, and more.
IoT is the concept of connecting any device to the Internet and to other connected devices (IBM, source).
Source: UN Department of Economic and Social Affairs (2018)
The rapid urbanisation will increase demand for services in urban areas exponentially and put pressure on population centres. In this future scenario, efficient, smart cities can represent a part of the solution.
Elkem has delivered metals and materials for the construction sector for several decades and play a key role in how cities are becoming better, smarter and more efficient.
In addition, silicones also have a wide range of usages within electronics.
“The extreme resistance of our materials, combining thermal and fire resistance as well as chemical stability, make silicones materials outstanding for long-term applications, where you either do not want to or cannot change materials frequently. This is the reason why silicones have become the material of choice in aviation, aerospace and automotive industry”, says Yves Giraud, global business manager in Elkem Silicones.
“For example, if you launch a satellite, you will not be able to change and inspect the materials every three years. The materials must be stable over a 15-year period in a very challenging environment. Another example is 5G antennas, which will become increasingly important as smart infrastructure, where Elkem’s material solutions are vital to protect critical functionalities and to reduce the need for maintenance and inspections for our customers”, says Giraud.
Another example is 5G antennas, which will become increasingly important as smart infrastructure, where Elkem’s material solutions are vital to protect critical functionalities and to reduce the need for maintenance and inspections for our customers”, says Giraud.
Reliable, sustainable and innovative
With increased demand for new energy solutions and smart applications, the role of cables is also becoming more important. To meet demand, manufacturers are looking for safer, more reliable, sustainable and innovative solutions.
Silicone rubber insulated cables provides both heat and fire resistance, and present high mechanical properties. The materials therefore contribute to protecting our lives in the cities.
“We believe smarter cities are one of several drivers that will increase the need for safe products that lasts. The use of silicones in smart application is a great reusable alternative, and is also of significant sustainability value, generating energy and saving CO2 emissions nine times greater than the impacts of production and recycling”, says Giraud.
“Standards are a hidden part of the information and communications technology networks and devices we all use daily”. This is how according to Chaesub Lee the technical standards have a vital role in achieving all SDGs. We all know that UN sustainable development goals are presently failing to have a meaningful impact; then this is perhaps a way out of this traumatic vicious circle.
Opinion: Technical standards have a key role in achieving the SDGs
Devex, 7 July 2022Technical standards can help address some of the most pressing needs of the planet. Photo by: Matthew Horwood / Alamy
Standards are a hidden part of the information and communications technology networks and devices that we all use every day. Though rarely perceived by users, they are vital in enabling the interconnection and interoperability of ICT equipment and devices manufactured by hundreds of thousands of different companies around the world.
For example, 95% of internet traffic is on fiber, built on standards from the International Telecommunication Union, a specialized agency of the United Nations for ICT. ITU has also played a leading role in managing the radio spectrum and developing globally applicable standards for 5G cellular networks.
But while technical standards are clearly indispensable for business and society to work in our industrialized world, it is also becoming clear that technical standards have a key role in addressing the Sustainable Development Goals.
Indeed, the focus of the recent ITU Global Standards Symposium, which brought together more than 700 industry leaders and policymakers, was how standards can help address some of the most pressing needs of the planet, such as eradicating poverty or hunger and mitigating climate change.
To address SDGs 1 and 2 on ending poverty and hunger, an ITU focus group on “Artificial Intelligence (AI) and Internet of Things (IoT) for Digital Agriculture” is working toward new standards to support global improvements in the precision and sustainability of farming techniques.
Under ITU and the World Health Organization, a focus group on “Artificial Intelligence for Health” aims to establish an “open code” benchmarking platform, highlighting the type of metrics that could help developers and health regulators certify future AI solutions in the same way as is done for medical equipment. Also, standards for medical-grade digital health devices — such as connected blood pressure cuffs, glucose monitors, or weight scales — are helping prevent and manage chronic conditions such as diabetes, high blood pressure, and heart disease.
Standards are helping bring broadband to rural communities with lightweight optical cable that can be deployed on the ground’s surface with minimal expense and environmental impact. The installation of ultrahigh-speed optical networks typically comes with a great deal of cost and complexity. Standards can change that equation by providing a solution able to be deployed at low cost with everyday tools.
To address SDG 11 on sustainable cities and communities, more than 150 cities around the world have started evaluating their progress toward smart-city objectives and alignment with the SDGs using so-called key performance indicators based on tech standards. These cities are supported by United for Smart Sustainable Cities, an initiative backed by ITU and 16 other U.N. partners.
International standards, recognized around the world, are essential for making technologies … accessible and useful to everyone, everywhere.
Addressing SDGs related to climate action and green energy, ITU standards for green ICT include sustainable power-feeding solutions for 5G networks, as well as smart energy solutions for telecom sites and data centers that prioritize the intake of power from renewable energy sources. They also cover the use of AI and big data to optimize data center energy efficiency and innovative techniques to reduce energy needs for data center cooling.
Financial inclusion is another key area of action to achieve SDG 1 on ending poverty. Digital channels are bringing life-changing financial services to millions of people for the very first time. Enormous advances have been made within the Financial Inclusion Global Initiative and the associated development of technical standards in support of secure financial applications and services, as well as reliable digital infrastructure and the resulting consumer trust that our money and digital identities are safe.
However, the complexity of global problems requires numerous organizations with different objectives and profiles to work toward common goals. Leading developers of international ICT standards need to work together to address the SDGs, using frameworks such as the World Standards Cooperation, with the support of mechanisms such as the Standards Programme Coordination Group — reviewing activities, identifying standards gaps and opportunities, and ensuring comprehensive standardization solutions to global challenges.
Including a greater variety of voices in standards discussions is crucial. It is particularly important that low- and middle-income countries are heard and that a multistakeholder approach is made a priority to have a successful and inclusive digital transformation.
Uncoordinated and noninclusive standardization can spell lasting harm for countries that already struggle to afford long-term socioeconomic investments. Without global and regional coordination, today’s digital revolution could produce uneven results, making it imperative that all standards bodies work cohesively.
Sustainable digital transformation requires political will. It was notable that last year in Italy for the first time, leaders from the G-20 group of nations used their final communiqué to acknowledge the importance of international consensus-based standards to digital transformation and sustainable development.
This important step could not have been made by one standards body alone.
Cities, governments, and companies face a significant learning curve while adopting new tech as part of low-carbon, sustainable, citizen-centric development strategies to meet the challenge of addressing the SDGs. International standards, recognized around the world, are essential for making technologies in areas like digital health and 5G — combined with bigger and better data use — accessible and useful to everyone, everywhere.
The views in this opinion piece do not necessarily reflect Devex’s editorial views.
.
About the author
Chaesub Lee
Chaesub Lee is the director of the Telecommunication Standardization Bureau at the International Telecommunication Union, a specialized agency of the United Nations for ICT. Lee has contributed to ICT standardization for over 30 years, specializing in areas such as integrated services digital networks, global information infrastructure, internet protocol, next-generation networks, internet protocol television, and cloud computing.
As of October 2021, 44 countries were reported to have their own national AI strategic plans, showing their willingness to forge ahead in the global AI race. These include emerging economies like China and India, which are leading the way in building national AI plans within the developing world.
Oxford Insights, a consultancy firm that advises organisations and governments on matters relating to digital transformation, has ranked the preparedness of 160 countries across the world when it comes to using AI in public services. The US ranks first in their 2021 Government AI Readiness Index, followed by Singapore and the UK.
Notably, the lowest-scoring regions in this index include much of the developing world, such as sub-Saharan Africa, the Carribean and Latin America, as well as some central and south Asian countries.
The developed world has an inevitable edge in making rapid progress in the AI revolution. With greater economic capacity, these wealthier countries are naturally best positioned to make large investments in the research and development needed for creating modern AI models.
In contrast, developing countries often have more urgent priorities, such as education, sanitation, healthcare and feeding the population, which override any significant investment in digital transformation. In this climate, AI could widen the digital divide that already exists between developed and developing countries.
The hidden costs of modern AI
AI is traditionally defined as “the science and engineering of making intelligent machines”. To solve problems and perform tasks, AI models generally look at past information and learn rules for making predictions based on unique patterns in the data.
AI is a broad term, comprising two main areas – machine learning and deep learning. While machine learning tends to be suitable when learning from smaller, well-organised datasets, deep learning algorithms are more suited to complex, real-world problems – for example, predicting respiratory diseases using chest X-ray images.
Crucially, neural networks are data hungry, often requiring millions of examples to learn how to perform a new task well. This means they require a complex infrastructure of data storage and modern computing hardware, compared to simpler machine learning models. Such large-scale computing infrastructure is generally unaffordable for developing nations.
Beyond the hefty price tag, another issue that disproportionately affects developing countries is the growing toll this kind of AI takes on the environment. For example, a contemporary neural network costs upwards of US$150,000 to train, and will create around 650kg of carbon emissions during training (comparable to a trans-American flight). Training a more advanced model can lead to roughly five times the total carbon emissions generated by an average car during its entire lifetime.
Developed countries have historically been the leading contributors to rising carbon emissions, but the burden of such emissions unfortunately lands most heavily on developing nations. The global south generally suffers disproportionate environmental crises, such as extreme weather, droughts, floods and pollution, in part because of its limited capacity to invest in climate action.
Developing countries also benefit the least from the advances in AI and all the good it can bring – including building resilience against natural disasters.
Using AI for good
While the developed world is making rapid technological progress, the developing world seems to be underrepresented in the AI revolution. And beyond inequitable growth, the developing world is likely bearing the brunt of the environmental consequences that modern AI models, mostly deployed in the developed world, create.
But it’s not all bad news. According to a 2020 study, AI can help achieve 79% of the targets within the sustainable development goals. For example, AI could be used to measure and predict the presence of contamination in water supplies, thereby improving water quality monitoring processes. This in turn could increase access to clean water in developing countries.
The benefits of AI in the global south could be vast – from improving sanitation to helping with education, to providing better medical care. These incremental changes could have significant flow-on effects. For example, improved sanitation and health services in developing countries could help avert outbreaks of disease.
But if we want to achieve the true value of “good AI”, equitable participation in the development and use of the technology is essential. This means the developed world needs to provide greater financial and technological support to the developing world in the AI revolution. This support will need to be more than short term, but it will create significant and lasting benefits for all.
GreenBiz came up with these six tips for deploying data-driven energy management to drive meaningful emission reductions through reducing building operating emissions at scale with data analytics. So here is a much down to earth way to a certain decarbonisation strategy.
Reducing building operating emissions at scale with data analytics
Nearly 40 percent of global CO2 emissions come from the built environment — with 28 percent resulting from buildings in operation. Whether your organization owns, operates or occupies a building, data-driven energy management is key to reducing its GHG footprint and Scope 1 and 2 emissions.
In the past, organizations have struggled to scale building operational energy improvement efforts for a variety of reasons. The most-cited reasons include organizational structures that fracture ownership of energy performance across disparate stakeholders, a lack of goal alignment and collaboration between landlords and occupiers, and the preponderance of legacy systems that make interoperability and data consolidation challenging.
According to United Nations projections, carbon emissions from buildings are expected to double by 2050 if action at scale doesn’t occur. With more companies pledging to decarbonize their business, and investors increasingly scrutinizing ESG data, scalable energy management will be a critical step in the transition to a low-carbon economy.
Today, we share six tips for deploying data-driven energy management at scale to drive meaningful emission reductions from your business.
Portfolio energy management software. Source: Envizi.
Collect meter-level energy consumption data where possible
Identifying GHG reduction opportunities should be a data-driven, systematic process. Start by examining building-level energy meter profiles and understanding how usage patterns relate to changing occupancy and weather conditions. Meters, which typically generate one datapoint every 15 to 30 minutes, as opposed to one datapoint every month or quarter on a utility bill, provide rich data to better inform your organization’s decarbonization strategy.
Tip: Leverage meter data, which provides real-time transparency of when and where energy is being used, to identify unexpected usage patterns and unlock higher-resolution benchmarking and analysis opportunities.
Benchmark the energy intensity of your building portfolio
Building-level energy management is powerful, but it never pays to operate in a vacuum. Understanding how a building performs compared to others provides context and can help your organization identify where to focus first. The approach to benchmarking depends on the type of buildings in your portfolio.
For example, typical portfolios of small to medium buildings (buildings of 4,000 to 20,000 square feet or so) often include many buildings dispersed across a geography (such as convenience stores, bank branches and fast-food stores), while large shopping centers, hospitals and universities manage larger, but fewer, centralized complex buildings.
Portfolios with larger commercial buildings can leverage third-party frameworks, such as Leadership in Energy and Environmental Design, Energy Star and NABERS, which compare energy intensity against an industry benchmark.
For portfolios of small to medium buildings that are dispersed, external benchmarks are harder to find. In this case, Envizi recommends internal benchmarking using meter data to make meaningful performance comparisons. Advanced normalization techniques can be applied to identify the poorest performers in the portfolio, which helps to inform a highly targeted strategy for improving efficiency and reducing emissions.
Tip: Undertake energy benchmarking before making investment decisions — don’t make the mistake of focusing on areas where there are no material savings. Envizi’s software can combine meter data with other contextual data (floor area, weather, operating schedules, and production units) to enable performance comparisons on a normalized basis.
Tune operational and behavioral efficiency
Buildings can be complex, but not as complex as building operations: the interaction between a building, its operators and occupants, and flow-on effects to energy performance.
Building services such as heating, ventilation and air conditioning (HVAC), which often account for almost 30 percent of annual emissions, are subject to continuous change and are often responsible for considerable “energy drift” over time due to poor operational practices. For this reason, technology that proactively informs and educates building operators is necessary to support time-poor operations teams to maintain optimum performance.
Often, manual audits will not detect the inefficiencies, but Envizi’s software uses a combination of continuous equipment monitoring, building management systems data, equipment nameplate data, weather data and other metrics to provide transparency to HVAC system performance and uncover operational issues that are otherwise difficult to detect.
Consider plant and equipment upgrades
Investing in equipment to deliver emissions reductions is dependent on an organization’s scale, scope and asset type and may be relevant only to building owners.
The appetite for plant and equipment upgrades may depend on how long the asset owner intends to hold the asset, the age of the building and the age of the equipment. Envizi recommends that building owners and operators engage their engineering consultants and specialist contractors to determine the feasibility of plant and equipment upgrades.
Tip: Technology can assist in the pre- and post-analysis of reduction projects to measure effectiveness and return on investment (ROI). Envizi’s software uses the International Performance Measurement and Verification Protocol to ensure calculations will withstand audit and validation.
Consider on-site and off-site renewables
After implementing solutions for operational, behavioral and system efficiencies, many organizations seek renewable energy as a proactive solution to get ahead on the decarbonization journey. Decisions on whether to procure on-site or off-site renewables are complex, and Envizi recommends coordinating with your organization’s engineering consultant or specialist contractor to assess its options.
Tip: Software platforms such as the one offered by Envizi can assist with monitoring the performance of solar assets, comparing the actual performance to promised performance and integrating the accounting of the renewable energy certificates to facilitate the most traceable reporting and auditing process.
Engage stakeholders
Energy management is rarely the remit of one team, but rather involves multiple stakeholders across an organization. The success of any emissions-reduction effort will be affected by the organization’s ability to effectively engage a cross-collaborative stakeholder group.
Typically, organizations with a strong culture of governance and executive ownership of the energy agenda can make the most impactful positive change. Often, inspirational leaders can make the difference with robust internal communication, empowerment through clear roles and responsibilities, and incentives for employees to take ownership of the energy reduction goals.
The transition to a low-carbon economy will require organizations to drastically increase the energy efficiency of buildings in operation. The following data-driven tactics can help your organization identify and achieve meaningful emission reductions:
Collect meter data where possible to understand granular energy consumption.
Benchmark the energy performance of the buildings by size/cohort in your organization’s portfolio to identify poor performers.
Use technology to monitor how HVAC systems are configured, to detect energy waste and optimization opportunities.
Before implementing equipment retrofits, solar photovoltaics or energy projects, engage a specialist to understand your organization’s options, and use data to establish a baseline against which to measure improvements.
Nominate a senior executive to champion your organization’s emissions-reduction program. A single system of record for emissions and energy can help enable cross-functional collaboration.
If you’d like to learn more about using data and technology to streamline and accelerate decarbonization, read “Pathway to Low-Carbon Guide.”
Generations of travelers have stood before the “ksars” of Djado, wandering their crenellated walls, watchtowers, secretive passages and wells, all of them testifying to a skilled but unknown hand.
Originally posted on DESERTIFICATION: Heidelberg Earth scientists study natural climate fluctuations of the past 500,000 years – https://www.labmanager.com/news/desertification-threatens-mediterranean-forests-30224 With a view towards predicting the consequences of human-made climate change for Mediterranean ecosystems, Earth scientists from Heidelberg University have studied natural climate and vegetation fluctuations of the past 500,000 years. Their primary focus was the effects…
Originally posted on HUMAN WRONGS WATCH: Human Wrongs Watch (UN News)* — Disinformation, hate speech and deadly attacks against journalists are threatening freedom of the press worldwide, UN Secretary-General António Guterres said on Tuesday [2 May 2023], calling for greater solidarity with the people who bring us the news. UN Photo/Mark Garten | File photo…
This site uses functional cookies and external scripts to improve your experience.
This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish.AcceptRead More
Privacy & Cookies Policy
Privacy Overview
This website uses cookies to improve your experience while you navigate through the website. Out of these cookies, the cookies that are categorized as necessary are stored on your browser as they are essential for the working of basic functionalities of the website. We also use third-party cookies that help us analyze and understand how you use this website. These cookies will be stored in your browser only with your consent. You also have the option to opt-out of these cookies. But opting out of some of these cookies may have an effect on your browsing experience.
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.
Privacy settings
Privacy Settings
This site uses functional cookies and external scripts to improve your experience. Which cookies and scripts are used and how they impact your visit is specified on the left. You may change your settings at any time. Your choices will not impact your visit.
NOTE: These settings will only apply to the browser and device you are currently using.
Google Analytics
To provide me with an idea of my site’s performance
