ESG
Beyond results, we are committed to help our customers meet their Environmental, Social and Governance goals. Sustainability and ESG are at the core of Sherpa.ai’s technology.
Our Platform helps ESG compliance
At Sherpa.ai we want to go beyond mere compliance and promote a responsible and ethical use of data as well as contribute to reduce the carbon footprint generated in model training.
We are committed to creating a positive impact in society and the environment.
Carbon footprint reduction
Digital Ethics
Green AI carbon footprint reduction
A large amount of energy and computational cost is required to train AI models in a centralized way, raising concerns about its carbon footprint. Total amount of energy consumed by data centres made up about 1% of global energy use over the past decade – equalling roughly 18 million US homes.
Green AI refers to AI that yields novel results while taking into account computational cost and environmental impact.
Organizations need to conduct rigorous exercises to determine the environmental impact and carbon footprint of, among other things, their model training, infrastructure and compute usage.
Federated Learning decentralizes data processing which reduces the number of communications and requires less energy consumption which decreases carbon footprint.
70%
Decrease of up to 70% in CO2 emissions with Federated Learning vs centralized training, according to University of Cambridge Research.
Researchers from the University of Cambridge have explored how Federated Learning can be more environmentally responsible than centralized learning.
See reports from Cambridge University:
- Can Federated Learning Save the Planet? Read the full report
- A first look into the carbon footprint of federated learning. Read the full report
Responsable AI Digital Ethics
AI‑powered solutions can sometimes be discriminatory, are unable to explain the decisions made by their algorithms, and potentially pose a risk to individual privacy given their heavy reliance on data. Issues related to AI development, such as “explainability”, transparency and accountability remain ongoing, raising questions about ethics, privacy and security.
Sherpa.ai’s Federated Learning platform promotes an ethical use of Artificial Intelligence as it allows organizations to extract all benefits of data and achieve highly accurate AI models without sharing any data. Protecting customer’s privacy and preventing bias in model training.
World Economic Forum ethical guidelines
Our solution is aligned with World Economic Forum’s guidelines of sustainability and ethical use of technology and data.
Ethics by Design: An organizational approach to responsible use of technology (WHITEPAPER – DECEMBER 2020)
Most companies today understand the importance of ensuring that the technology they employ is trustworthy (i.e. that it addresses foundational security, privacy and regulatory concerns).
While critical ethical thinking about technology may be a new skill set for some, almost everyone agrees that issues such as data privacy and algorithmic bias can pose significant reputational and financial risks if unaddressed.
Measuring Stakeholder Capitalism: Towards Common Metrics and Consistent Reporting of Sustainable Value Creation (WHITEPAPER – SEPTEMBER 2020)
A key principle for good governance is the effective oversight of corporate decision-making to ensure compliance with relevant laws and regulations, as well as meeting stakeholder expectations for ethical behaviour.
Data stewardship is also a critical area. As outlined by the World Economic Forum’s recent report, Integrated Corporate Governance, data stewardship priorities may include “cybersecurity, the use and governance of artificial intelligence and machine learning, and privacy and data ownership issues associated with data collection, management and use.”
Other benefits of Federated Learning
Sherpa.ai Federated Learning platform addresses other key infrastructure challenges, such as Bandwith Limitation in situations of unreliable connectivity (i.e. offshore facilities, remote locations) as well as Excess Latency or Network Congestions.