Disaster Recovery Journal Fall 2023

ties for data center companies to expand for profit and market share gain, if they care about the environment (and the harmful gases data centers produce in the atmosphere), one of the possible avenues which may be implemented to reduce carbon emissions to support sustainable AI would be to consider the backup power which uses diesel fuel to run systems. However, diesel fuel is refined from crude oil and produces harmful pollutants. Some petroleum industries are producing ultra low sulfur diesel (ULSD) fuel which is cleaner and contains less sulfur. In 2021, U.S. Energy Information Administration estimated 26% of the U.S. transportation released approximately 10% of total U.S. energy-related CO2 emissions, but there will be improvements with ULSD fuel. Clean Energy and Net Zero The continued effort to adopt AI throughout the industry is now causing companies to research a replacement of diesel fuel to power their backup data cen ters in the continued effort to reduce the footprint of CO2 emissions in the environ ment. While hydrogen stationery power in data centers does not meet the challenges of diesel vehicles, it could contribute to the reduction of harmful gases in the envi ronment. AI applications which depend on large datasets running on powerful systems in data centers have a reliance on backup sys tems deemed as important as production systems. In fact, backup systems could be considered production, especially if data usage is bi-directional between systems and data centers. The use of hydrogen power as a replace ment to diesel fuel can provide clean energy which is reliable and readily avail able. But with every positive advance ment, there are challenges and sometimes downside effects. For instance, the expan sion of AI and the dependency on data is driving the need for more data centers and the adoption of more clean energy such as wind turbines, hydro fuel, and solar energy. The hydro fuel market appears to be dominating the market, requiring heavy reliance on water. As such, water is becoming a commodity and available for

sale by companies that have established themselves as sustainable water resource companies. Water-based companies are now clos ing off access to water and turning water into a business by selling or leasing usage to farmers, other companies, and indus tries. With the advancement in AI and water becoming a commodity, it will become more expensive to run data cen ters with hydro fuel. Companies are continually searching for water friendly regions suitable for data centers and will have to accelerate their research to claim water supply when dis covered. The race to net zero with clean energy is good for the environment, but companies should be mindful of citizens in the region who may be impacted by the unavailability of water. Some communi ties rely heavily on available clean, free water used for farming and necessary to sustain life. Once companies block water and sell clean energy for data centers, the community suffers. This is an area requir ing additional study to help understand if a balance between citizens, for-profit companies, and data center owners can be achieved. Collaboration and Governance Sustainable AI is a global concern, requiring the involvement of the gov ernment and private companies to come together to discuss, agree, and implement guidelines and policies to address devel opment in AI that is proportionate to the environment and people’s rights. The reg ulation of the EU Parliament and the EU Council published a 2021 paper on AI from an ethical standpoint and highlighted the risks associated and the development of an ecosystem of trust. The importance of sus tainable AI led to the council establishing a European Artificial Intelligence Board, tasked with the responsibility of providing guidance on emerging AI issues and assist ing in ensuring regulatory requirements are adhered to within the industry. In addition, the US-EU Trade and Technology Council inaugural joint statement provided results from a study exploring the impact of AI on wages and loss of employment. It would be helpful if the US-EU commission con

ducted a study on sustainable AI from the standpoint of environmental impact and elevate the importance of sustainability within the industry. Conclusion To provide a more sustainable AI envi ronment, the public and private sectors must do their part to ensure advances in AI can be accomplished with clean energy, which will not harm the environment and add to the carbon footprint. One of the goals for sustainable AI should be to pro tect the rights of citizens’ health as more water is consumed for data centers. Data center owners are researching possible solutions to reduce their carbon footprint by utilizing the power from wind turbine, solar, and hydro energy. They should look at the downside caused by developing and implementing this alternative energy and provide opportunities to implement envi ronmental sustainability for the future of AI. Policy makers appear to be involved in AI from an ethical and moral perspec tive and are writing policies to help pro tect the privacy of citizens and workforce employment. The advancement in AI will continue. It is important that citizens’ con cerns and the effect on the environment are considered early on in budget planning and development for a more sustainable technological society. v

Brian Stanislaus has extensive knowledge in business resilience from his past roles from developer to senior vice president of global resiliency. Stanislaus is presently responsible for business resiliency at New

York Power and Authority, ensuring mission critical func tions at the sites and canals are available in the event of a disruption affecting the company’s ability to provide carbon-free, affordable, clean, and reliable electricity. Stanislaus is a former U.S. Marine and active member in a veteran organization. He has a master’s degree in infor mation assurance and is a doctoral candidate in national security with significance in artificial intelligence.

Chandregowda (Chandra) Pachegowda, a seasoned IT professional, has held diverse digital roles across Asia, Europe, and North America, assisting clients in their digital trans formation endeavors. Presently, he serves

as a director of cloud architecture at S&P Global Inc. Pachegowda possesses a master’s degree in information systems with a cybersecurity certificate, and he is currently pursuing a doctorate of science degree specializing in cybersecurity, social engineering, and AI ethics.

36 DISASTER RECOVERY JOURNAL | FALL 2023