The field of artificial intelligence is shifting towards more sustainable and energy-efficient solutions. Researchers are exploring new approaches to reduce the environmental impact of AI systems, including the development of domain-specific models, energy-aware frameworks, and human-AI collaboration. The use of analog computers and reconfigurable systems is also being revisited as a potential means to reduce energy consumption. Furthermore, the environmental impact of continuous integration and continuous deployment (CI/CD) pipelines is being investigated, highlighting the need for more sustainable software development practices. Noteworthy papers include: Energy-Efficient Domain-Specific Artificial Intelligence Models and Agents, which introduces a novel framework for sustainable AI solutions, and Toward Carbon-Neutral Human AI, which advocates for a shift towards human-inspired, sustainable AI solutions. Additionally, Fuzz Smarter, Not Harder presents an energy-aware framework for greener fuzzing, and Environmental Impact of CI/CD Pipelines investigates the carbon and water footprints of GitHub Actions.
Sustainable AI and Computing Systems
Sources
Lincoln AI Computing Survey (LAICS) and Trends
Privacy by Design: Aligning GDPR and Software Engineering Specifications with a Requirements Engineering Approach
HW/SW Co-design of a PCM/PWM converter: a System Level Approach based in the SpecC Methodology
Energy-Efficient Domain-Specific Artificial Intelligence Models and Agents: Pathways and Paradigms
Toward Agents That Reason About Their Computation
Toward Carbon-Neutral Human AI: Rethinking Data, Computation, and Learning Paradigms for Sustainable Intelligence
Towards Human-AI Synergy in Requirements Engineering: A Framework and Preliminary Study
Fuzz Smarter, Not Harder: Towards Greener Fuzzing with GreenAFL
An Agentic Framework for Rapid Deployment of Edge AI Solutions in Industry 5.0
Reconfigurable Analog Computers
Environmental Impact of CI/CD Pipelines