ABSTRACT
This paper explores the intersection of sustainability, freedom, and financial data science, focusing on how augmented intelligence (AI) can support society’s need to be “free from” harmful activities like climate change while preserving the “freedom to” pursue sustainable and competitive economic choices. Drawing from Isaiah Berlin’s concepts of negative and positive freedom, the session reframes sustainability as a balance between avoiding harm and enabling opportunity. It emphasizes the urgency of accurate, precautionary CO₂e data, particularly where regulatory pressure is high and corporate disclosure is weak. The speakers present a Financial Data Science (FDS) pipeline powered by agentic augmented intelligence, capable of generating realtime, high-quality CO₂e estimates across all scopes and categories. This technology supports more credible, actionable, and outcome-driven sustainable finance by empowering institutions with data that reflect the true environmental impact of their investments—regardless of self-reporting biases. Ultimately, the paper advocates for AI systems that are transparent, accountable, and human-supervised, supporting both effective regulation and investor decision-making.
Keywords: sustainability, freedom, financial data science, augmented intelligence, AI, CO₂e emissions, Scope 3, ESG data, climate risk, precautionary principle, regulatory compliance, negative freedom, positive freedom, sustainable finance, impact measurement, data accuracy, investor decision-making
Content
Sustainable finance can support society to meet its objectives by channelling capital into sustainable economic activities. However, there remains a lack of clarity regarding the extent to which sustainability shall embrace or curtail freedom. In «Two Concepts of Liberty», Isaiah Berlin (2016) defines negative freedom as the ‘freedom
from’ interference, meaning the absence of something. Positive freedom, on the other hand, is the ‘freedom to’ choose and shape one’s life, meaning the presence of something (Berlin, 2016). We relate Berlin’s concepts of freedom to sustainability through the definition provided by the Brundtland Commission Report, in which
sustainability means meeting the needs of the present generation without compromising the ability of future generations to meet their own needs (WCED, 1987). In the ‘freedom from’ perspective, harms should be avoided so that current and future generations can meet their basis needs (Wood, 2023). For example, fossil fuelinduced climate change should be avoided, as it jeopardizes the food availability of both present and future generations. In the ‘freedom to’ perspective, people have the freedom to define, choose, and pursue their own goals (Sen, 2024). For instance, individuals may choose to specialize in impact management. We relate both freedom
perspectives to three types of activities – harmful, neutral, and helpful – while acknowledging the basic need of every economy to be competitive (Draghi, 2025). As shown in Figure 1, the freedom to invest in helpful or neutral activities without material differentiation appears crucial for competitiveness while the freedom from the harmful causes of climate change is crucial for achieving the Paris Agreement.
This conceptualisation of sustainability and freedom implies that most pressing need for impactful, accurate data delivered by a custom Augmented Intelligence (AI) financial data science is the quantification of harmful activities where regulation is likely and discretion is low. In other words, the substantial discretion likely awarded to helpful or neutral activities reduces the need for accuracy as much more variation is tolerable. While society’s freedom from several harmful activities is crucial, climate change and related or unrelated wars are probably the most severe risks of our time. For exemplary purposes, this note focuses on the former to demonstrate the Impact
that can be achieved via Augmented Intelligence (AI) and Financial Data Science.
With respect to climate change, accurate and instantly updated CO2e emission data that emphasises society’s need to be free from climate change still appears an illusion for three reasons. First, to date a relevant number of corporations in public or especially private capital markets do not report CO2e emissions, not even Scope 1 and 2. Second, despite regulation in the EU being very clear that estimations should use the Precautionary Principle (aka “if in doubt, err on the side of the planet”), too many estimation models in use today are simplistic point estimates that will underestimate CO2e emissions as soon as the majority of self-reported input data is underreported, which is usually the case. It is crucial to note that accuracy viewed from a freedom from perspective is not the most likely estimate but the estimate which is with a sufficient likelihood (e.g. 90%) not too low. Such a Precautionary Principle based CO2e estimation process in line with EU regulation is displayed in
Figure 2.
The third reason why accurate and instantly updated CO2e emission data remains illusive lies in insufficient use of agentic augmented intelligence driven financial data science (FDS) processes which achieve both, a very high (e.g. six sigma level) quality standard and a delivery of CO2e data updates within a few weeks rather than many months. Crucially, the (FDS) process has to represent augmented rather than artificial intelligence because (i) artificial models can only learn do execute without error tasks which virtually any human in the training data can do without error and (ii) accuracy requires ownership and responsibility which to date can only be assumed by humans. In other words, given insufficient clean (i.e. error free) training data on complex CO2e disclosure aspects where the reporting companies themselves can err the fully artificial process is unable to train accurately and, complementarily, humans remain needed to take ownership and thereby blame in today’s compliance regimes.
Using an augmented intelligence driven financial data science pipeline with four AI agents supervised by financial data scientists and data quality controlled by climate data scientists, we can estimate precautionary live CO2e data for every Scope and category for thousands of listed and unlisted companies at any precautionary threshold (e.g. 90%, 95%, lower or even higher). Such technological ability allows for an array of new analytical insights such as (i) which Scope 3 categories are the most impactful, (ii) which sectors are most material in each category or (iii) which firm would have the highest emissions if everyone was monitored (regardless of self-reporting or not). Figure 3 provides indicates answers to these three questions using a universe of about fifty developed and emerging economies.
Authors
- Andreas Hoepner, Professor of Operational Risk, Banking & Finance at University College Dublin (UCD). Co-Inventor EU Paris-Aligned Benchmarks & EU Taxonomy Aligning Benchmarks.
- Florian Faust, Software Engineer, Sociovestix Labs.
Furthermore, this document is signed in a personal capacity and does not represent the official position of the institutions or entities to which the author may belong.
Oxford/25 Congress Final Report
Reaching Pragmatism in Sustainability
#Impact #Engagement #Megatrends #Data powered by AI
This comprehensive report defends that sustainability can no longer rest on labels or narratives alone. It must be anchored in credible transition plans, robust data, coherent regulation and real-world outcomes.. Dive into the findings and help shape a sustainable future.
