Building Ethical and Equitable AI: Addressing Power, Bias, and Environmental Impact
AI-generated context summary requested by a Free News Reader user. Sourced via Gemini from publicly available information — no paywalled content was accessed.
You hit a paywall. Here’s the context on this topic based on publicly available information. We did not access any paywalled content. View original article.
Building Ethical and Equitable AI: Addressing Power, Bias, and Environmental Impact
- The Encode: AI for Science Fellowship, launched in February 2025 by Pillar VC, aims to bridge the gap between AI talent and scientific research to address global challenges.
- Leah Morris, Executive Director of the Encode: AI for Science Fellowship, emphasizes the importance of integrating industry and academic expertise to commercialize scientific research.
Full Summary — powered by AI
The rapid advancement of artificial intelligence presents both immense potential and significant challenges, particularly concerning power dynamics, inherent biases, and environmental impact. Leah Morris, Executive Director of the Encode: AI for Science Fellowship, highlights the critical need to “build what matters” by aligning AI development with human well-being.
One major concern is the substantial energy consumption of AI. Data centers, crucial for training and deploying AI models, are projected to consume approximately 1.5% of global electricity in 2024, with a 12% annual growth rate over the past five years. AI-focused data centers are particularly energy-intensive, using around 60 kilowatts of electricity per server rack compared to 10 kilowatts for general-purpose data centers. This surge in demand strains local grids and can lead to construction delays for new data centers due to power availability bottlenecks and supply chain shortages for critical equipment like transformers. Beyond electricity, these facilities also require significant amounts of water for cooling.
Another pressing issue is the potential for AI to exacerbate existing inequalities through bias and discrimination. Biases can be introduced through human perspectives, data collection practices, and cultural contexts, leading to unfair treatment and unequal opportunities. The lack of diversity among AI development teams can also contribute to these biases. Ensuring fairness requires training AI models on diverse datasets and incorporating inclusive design practices.
The Encode: AI for Science Fellowship, a one-year program launched by Pillar VC in February 2025 with support from the Advanced Research + Invention Agency (ARIA), seeks to address these challenges by embedding top AI talent in leading UK research labs. The program aims to accelerate scientific breakthroughs in areas such as neurotechnologies, climate and energy, robotics, materials science, and biological sciences. The inaugural cohort of 18 fellows, announced in October 2025, received £8 million in government funding to apply AI to ambitious research initiatives, including cancer prevention and biomaterials design. The fellowship emphasizes translating frontier research into real-world solutions through startups, open-source communities, or research-driven non-profits.