Kunstig intelligens
Kunstig intelligens har et enormt potentiale til at bidrage til den grønne omstilling, men teknologien er også i sig selv lidt af en klimasynder. Det paradoks, samt en række andre udfordringer ved brugen af kunstig intelligens, ser vi nærmere på i vores forskningsnetværk "AI and green transition"
Kunstig intelligens er et tveægget sværd for den grønne omstilling.
På den ene side har teknologien et enormt potentiale til at effektivisere vores brug af ressourcer, styrke måden vi træffer beslutninger, optimere vores brug af vedvarende energikilder og skabe grønne jobs.
På den anden side er kunstig intelligens også selv en relativt stor belastning for klimaet. Udvikling og implementering af AI-systemer kræver betydelige mængder energi. Dette kan i sig selv resultere i øgede udledninger af drivhusgasser og bidrage til klimaforandringer. Dertil kommer problemer med bias og diskrimination, privatlivs- og sikkerhedsproblemer samt andre etiske udfordringer.
Vores forskningsnetværk "AI and green transition" arbejder med at vurdere potentialer og risici ved brug af kunstig intelligens i den grønne omstilling.
Kunstig intelligens som grøn løsning
Forskere i netværket
Raghavendra Selvan | Department of Computer Science | Faculty of Science |
Christian Igel | Department of Computer Science | Faculty of Science |
Yevgeny Seldin | Department of Computer Science | Faculty of Science |
Trine Krogh Boomsma | Department of Mathematical Sciences | Faculty of Science |
Olga Kokoulina | Centre for Information and Innovation Law (CIIR) | Faculty of Law |
Beatriz Martinez Romera | Centre for International Law and Governance | Faculty of Law |
Læs hele beskrivelsen (på engelsk)
Aim
Recent advancements in Machine Learning (ML) and Artificial Intelligence (AI) methods can play a significant role in supporting the green transition of societies. AI methods can assist in decision-making and enable more efficient and effective use of resources, reduce emissions, optimize the use of renewable energy sources and enable green jobs in the employment market. While AI has the potential to support the green transition of societies, it is also important to consider its environmental impact as the development and deployment of AI systems require significant amounts of energy. This can result in increased greenhouse gas emissions and contribute to climate change. Further, it is important to consider concerns about bias and discrimination, privacy and security issues, and other ethical implications when developing and deploying AI systems. The aim of this thematic solution will be to collaboratively assess the potentials and risks of using AI methods in pursuit of the UN Sustainable Development Goals.
Description
ML/AI methods have made significant progress in recent years and have been applied to a wide range of domains, including healthcare, finance, transportation, entertainment, and more. By leveraging advanced ML algorithms and big data, AI methods can support more sustainable and efficient practices across a range of scenarios.
The green energy transition involves decision-making at all stages of the process and in all parts of the system. It is of vital importance to develop models and methods that facilitate quantitative analysis and provide decision support tools for control, planning and assessment. One way AI methods can contribute to the green transition is by enabling the integration of renewable energy sources like wind and solar power into the energy grid. AI algorithms can optimize the storage and distribution of renewable energy, ensuring that it is used efficiently and effectively. Another way AI methods can also be used is to monitor and analyze environmental data, allowing for early detection and response to environmental risks like pollution, and natural disasters. For example, AI methods can be used to analyze satellite images to track deforestation or estimate carbon storage in trees. AI algorithms can also be used to optimize energy usage and reduce waste in buildings, transportation, and manufacturing. AI methods can analyze data on energy consumption and identify patterns and inefficiencies, allowing for targeted interventions to reduce energy consumption and greenhouse gas emissions. Finally, AI methods combined with behavioural science and psychological theory, can be used to better understand the drivers of consumption behaviours captured in digital traces.
As the use of AI methods continues to expand, it is also important to consider its potential environmental impact. The development and deployment of AI systems require significant amounts of energy and resources, which can result in increased greenhouse gas emissions and contribute to climate change. For example, the training and operation of AI models require large amounts of energy (to run computation on data centers). These systems consume large amounts of electricity, which is often generated from non-renewable sources like coal and natural gas. Additionally, the production of AI hardware and equipment requires significant amounts of resources and energy, as well as the extraction of raw materials like metals (embodied emissions).
To mitigate the environmental impact of AI, it is important to reduce the carbon footprint of AI hardware and software. Sustainable AI architectures and algorithms that can achieve similar results with less energy consumption can also be developed. Recycling and proper disposal of e-waste can help reduce the environmental impact of AI hardware and equipment.
The use of AI systems to aid the green transition shall be aligned with the emerging regulatory requirements of their responsible and ethical deployment. The enforcement of this commitment requires the collaboration of various stakeholders, including governments, industry leaders, and civil society groups. The legal framework comprises international, regional and national norms, enforced through both formal sanctions as well as leniency programmes. To unleash the potential of AI to bring significant benefits to the green transition of societies, legal, ethical, and other societal considerations and concerns should be articulated and tackled in a future-proof manner.
Overall, AI has the potential to bring significant benefits to the green transition of societies. By optimizing energy usage, integrating renewable energy sources, and monitoring environmental risks, AI can help reduce greenhouse gas emissions and support the transition to a more sustainable future. This thematic solution which comprises experts from law, social sciences and ML/AI is well positioned to collaborate and use synergy to develop a coherent framework that will proactively tackle the outlined challenges.
The thematic solution is established in collaboration with UCPH SCIENCE AI Centre.