For nearly three decades, the world has annually witnessed an event of critical importance for the future of our climate: the Conferences of the Parties, better known as the COP. First held in 1995 following the adoption of the United Nations Framework Convention on Climate Change at the Earth Summit in Rio in 1992, these conferences bring together nations that have ratified this treaty in a collective effort to combat climate change, a phenomenon increasingly evident in our daily lives.

Among these conferences, two stand out for their significant impact. The first COP3, held in Kyoto in 1997, marked a turning point with the near-unanimous adoption of the Kyoto Protocol. This agreement, which came into force in 2005 after intense negotiations, mandated signatories to reduce greenhouse gas emissions by at least 5% by 2012. Despite its legally binding nature, some countries attempted to diminish its ambition, and others, like the United States, never ratified it. Canada withdrew from the treaty in 2011, citing the discovery of highly polluting tar sands in Alberta. At the 2012 Doha conference, the Kyoto Protocol was extended until 2020 despite the absence of the US agreement.

COP15, held in Copenhagen in 2009, acknowledged for the first time the necessity of limiting global warming to 2°C above pre-industrial levels and proposed the creation of a Green Climate Fund endowed with 100 billion US dollars annually until 2020. Unfortunately, this initiative lacked legal enforcement and clear rules for fund allocation. By 2014, after the Lima conference, the Green Climate Fund had only amassed 10 billion US dollars.

Then came COP21 in 2015 in Paris, one of the most well-known conferences, which led to the landmark Paris Climate Agreement. This agreement set three primary goals:

1. Limit Temperature Rise: Keep the global temperature rise well below 2 degrees Celsius above pre-industrial levels while pursuing efforts to limit it to 1.5 degrees Celsius.
2. Adapt to Climate Impacts: Enhance the ability of countries to adapt to climate change impacts, focusing on resilience and adaptive capacity, especially in vulnerable regions.
   
3. Align Financial Flows: Redirect financial flows towards low greenhouse gas emissions and climate-resilient development, ensuring consistent support for mitigation and adaptation.

Once again, the agreement was not, or only minimally, binding: Participant countries were encouraged to define their "Nationally Determined Contributions" to be re-evaluated and submitted to the UN every five years, with each submission expected to be more ambitious than the last. The only legal obligation was the transparency of national contributions and their evaluation by experts.

The Paris Agreement, however, paved the way for landmark climate litigation, including a significant case in the Netherlands where a foundation sued the Dutch government for reducing its climate ambitions. The government lost, with the European Convention on Human Rights forming the legal basis of the decision.

From Words to Deeds: The Struggle for Effective Climate Change Policies at COP28

The climate is a highly complex system with significant inertia; actions taken today might only manifest their effects in a century! As of 2020, global warming is estimated to be around +1.2°C, with an increase of approximately +0.2°C per decade.

Current policies are steering us toward a +3°C increase, underscoring the need for a COP that results in a binding agreement backed by major powers and supported by financial measures. Former UN Secretary-General Ban Ki-Moon had suggested a global tax on financial transactions to fund the Green Climate Fund.

There is also hope for new agreements to ban subsidies for fossil fuels. However, the fact that COP28 is set to take place in the United Arab Emirates, chaired by the CEO of the national oil company, sends a mixed message. It is crucial that Gulf countries play a significant role on the international stage, especially given the recent escalation of the Israeli-Palestinian conflict in late 2023, highlighting their pivotal role in both international peace and climate change issues. On the latter, the trajectory is concerning: 181 million tons of oil were extracted in 2022, an increase of nearly 11% in a year, and gas extractions, though stable over the past year, have risen by 9% since 2012.
COP28, therefore, faces legitimate criticism, with the most significant being that the conference could be an exercise in greenwashing. Recent allegations reported by the BBC suggest potential misuse of the COP presidency to secure new oil and gas contracts.

Finally, responsibility contributions remain unresolved: the "Economic North" is primarily responsible for climate change, yet those who will suffer the most are the countries of the "Global South." Some island nations are even at risk of disappearing due to rising sea levels caused by climate change, and certain areas could become uninhabitable by 2050 due to extreme temperatures and humidity, preventing natural cooling processes like sweating. Addressing loss and damage will also be a central point at the conference.

Stay tuned for the second part summarizing the debates and agreements done at COP 28.

Online, Japan Investor Forum, July 2022.

Sylvain Forté, SESAMm's co-founder and CEO, discusses ESG data and its challenges. Further, he describes how to generate insights and reports on millions of companies, including micro-companies, using artificial intelligence and natural language processing.

Below is an approximation of this video’s audio content. Watch the video for a better view of graphs, charts, graphics, images, and quotes the presenter might be referring to in context.

About SESAMm

To give you a bit of context, I’m CEO of SESAMm, a French company of around 100 people that has been in business for eight years and that specializes in artificial intelligence for finance, especially with a focus on ESG.

So we work with some of the largest insurance companies in Japan, such as Tokio Marine, Asset Management One, or Japan Post Insurance. And we have seen the rise of ESG investing in the past few years, especially in the past four years in Europe and in the U.S. And we see now this trend also in Asia and in Japan, more specifically.

Primary uses of ESG data

The primary uses of ESG that we see are first complying with regulation. That is the key priority for most asset managers, but also improving performance. Many quantitative teams are seeing ESG also as a way to have new factors integrated that could qualify to generate alpha in investment funds. ESG is also used a lot in order to better manage risk in portfolio and, finally, to better analyze sustainable investment opportunities.

ESG use cases

So a couple of the main use cases are detecting ESC controversies. So purely from the perspective of generating risk alerts, excluding assets that are not well rated in portfolios, or creating portfolios that contain best-in-class assets, meaning most sustainable assets.

And finally, I want to mention that this trend is really global. So it's across both public assets, equities, and bonds, and also across private equity. And we see private equity reacting very quickly to the ESG trend.

Traditional ESG data challenges

So now, let's discuss in more detail some of the key challenges of ESG data. Traditionally, ESG data is created by teams of analysts that are looking at individual companies that are gathering data from each of the companies, and that are then reading the press in order to complement that information. This approach is relevant, but it is hard to scale, and it presents some difficulty. Traditional ESG ratings agencies are, for example, MSCI or system analytics.

The problem with a lot of traditional ratings is that they don't cover small companies very well. And this is one of the key challenges currently in ESG is the lack of coverage. So it is very difficult to cover small caps, microcaps, and also private companies. In particular, in Asia, the coverage is very poor right now for ESG, and that means that many portfolio companies may not be covered by ESG rating. In Japan specifically, even large companies are sometimes not covered by traditional ESG providers. So that creates a lot of data inefficiency in the industry.

Another key challenge that we see in ESG right now is the frequency of ESG ratings. So oftentimes, ESG ratings are updated only one time per year or just a few times per year. And when ESG ratings are used for risk management, obviously, the market is moving much more quickly than one time or a few times per year.

In addition to that, we see that ESG ratings mostly takes into account information that is reported by management and does not take as much into account information that is from outside of the company. For example, in the case of government scandals, such as fraud scandals, it is actually better to have information that is not reported by the company but that also has an external point of view.

Lastly, the last key challenge I want to mention in ESG data specifically, and one challenge that I'm sure you are aware of in market data and fundamental data is that ESG data is oftentime, not point-in-time. So that means that you don't have a continuous dataset that has not been modified over time. ESG agencies tend to modify their ratings after the fact, and so that means that the rating that you will receive now for a data point in 2020 will not be the same that the rating that you would actually have received in 2020 point-in-time. That creates a lot of problems when you want to back-test data because you cannot reproduce actual historical results.

So these are all of the key challenges that we have identified in ESG data currently, and there are challenges in order to address the needs that we described. But there are actually some solutions that exist.

The solution to ESG data challenges

And one of the key solutions right now that is merging in ESG is the use of artificial intelligence, in particular, what is called natural language processing, meaning text analysis.

What we do at SESAMm and what some other providers do is detecting ESG risks and positive impact with regards to sustainability by analyzing automatically billions of articles and messages in real time. So as an example, we have 18 billion articles and messages from common news websites, from social media, from blogs and forums, and from company reports. And we automatically detect ESG themes and risk and perform sentiment analysis in order to understand whether a company may be exposed to an ESG controversy or whether a company may have positive impact with regards to sustainability.

Advantages of AI for ESG data challenges

And the advantage of AI in that context is that it solves a lot of the challenges that we discussed before. So it helps access higher frequency data, it helps cover small companies, private companies, it helps also find information that is independent, that is public, and that is not necessarily just reported by management, and it also is point-in-time information that can easily be backlisted.

How SESAMm tackles ESG data challenges

So I'll mention a couple of use cases to illustrate that in more detail. But basically, at SESAMm, we create an ESG datasets in order to track more than 90 different ESG risks and also the 17 sustainable development goals in order to precisely identify positive impact. And we do that on millions of companies, not just large public companies but also small caps and also private companies.

SESAMm ESG data use cases

Some of the use cases that I wanted to illustrate for that is using artificial intelligence in order to perform ESG monitoring using alerts. What that means is that we automatically generate ESG alerts on portfolios, for example, of equities or bonds on a daily basis, including portfolios of Japanese equities. And this data is then used by quantitative analysts and also fundamental managers to systematically exclude companies that are exposed to controversies in a portfolio. And this is a very efficient approach to systematically exclude companies that are not sustainable that are exposed to them.

Secondly, we have companies generate ESG signals by combining market data and ESG AI data to generate alpha. So basically, we create long-only and long-term portfolios, and we incorporate these ESG signals in order to improve the alpha of these portfolios.

The two last examples I wanted to mention, one is positive impact. So there is a specific framework called the UNSDGs for sustainable development goals, which is well suited to automatically detecting positive impact actions by a company, such as implementing, for example, a new net zero carbon policy. And we automatically track these announcements and these positive actions that companies perform in order, again, to share this information in the form of alerts to help fundamental managers track the sustainability actions of their portfolio companies and automatically report on them without having to do manual research.

The last use case I wanted to illustrate, and it's going to be my last point, is due diligence in private equity. So this is not only applicable to public assets but also to private assets. As an example, we have the Carlyle Group, a very large private equity company in particular with the Japanese team, and we have them generate various kinds of analytics at the stage when they evaluate the company. And in particular, we help them monitor and track potential ESG risk and sustainability factors which are very important to assess potential private assets opportunities. So this is the last use case that I want to mention. And as you can see, there are many opportunities in a growing field in ESG that started in Europe and came out to Asia. But there are also a lot of the challenges which artificial intelligence can help solve in some cases and which are illustrated with some examples.

Thank you very much.

A familiar debate has followed ESG data for years. One camp argues that the field generates too much information for any human team to handle, so the work should be left to machines. The other argues that ESG judgments are too consequential to automate, so humans should review everything. Both positions contain a real concern. Neither describes how a credible rating is actually produced.

With the publication of its full Controversy Exposure Score methodology, now public and free to access following the entry into force of the EU ESG Rating Regulation on 2 July 2026, SESAMm is making the answer explicit. A trustworthy rating is not a choice between artificial intelligence and human expertise. It is the disciplined combination of the two, with each doing the part of the work it does best.

The Scale Problem Is Real

Teams that monitor ESG controversies rarely suffer from too little information. They suffer from too much. A single incident can generate dozens of articles within days, in multiple languages, across outlets of very different quality. Multiply that by a global investment universe and the volume becomes impossible to track by hand.

This is the part of the problem that machines are built for. SESAMm's pipeline ingests more than 10 million documents a day, drawn from an input layer of over 30 billion documents that includes licensed global news, public web and media feeds, NGO publications, and public regulatory and judicial filings. It screens controversies for millions of public and private companies, alongside infrastructure projects, state-owned entities and sovereigns. No analyst team could read at that scale, and none should try. Asking people to do machine work is how important signals get missed.

So artificial intelligence carries the load. Natural language processing and machine learning models, including large language models, attribute documents to the right entity, filter for genuine ESG relevance, and group related articles into discrete events and events into continuous cases. This is what allows a controversy that unfolds over weeks to be tracked as one developing story rather than a hundred disconnected headlines.

Why Scale Alone Is Not Trust

A system that reads everything will also, inevitably, misread some of it. SESAMm is direct about this in its methodology, because pretending otherwise would be the opposite of transparency.

Probabilistic language models can misinterpret a historical or hypothetical reference as an active controversy. Automated clustering can occasionally merge two distinct incidents or split one prolonged crisis into fragments. Model accuracy varies across languages, and lower-resource languages or heavily idiomatic content raise the risk of misclassification. These are structural properties of statistical systems, not bugs to be wished away.

This is precisely where scale stops being enough and human expertise becomes indispensable. A number that informs how capital is allocated cannot rest on automation alone.

Where Human Judgment Enters

SESAMm operates a dual-layer human quality-assurance process, and it runs every day.

At the first layer, a dedicated Research and Analytics quality-assurance team reviews data accuracy, both reactively, when a question is raised about a case, a score or a classification, and proactively, by reviewing generated alerts. Where an issue is confirmed, the correction, whether a reattributed entity, a corrected sub-risk tag or the removal of an irrelevant event, is applied at the source, logged, and the affected scores are recomputed on the standard daily cycle.

At the second layer, complex cases and recurring structural issues are escalated to the Methodology Lead, who can update the underlying training corpus so that a category of error becomes less likely in future. This is the detail that matters most. Human review is not a final rubber stamp on top of the machine. It is a feedback loop that teaches the system, so that today's corrections improve tomorrow's automated output.

The same expertise sits at the front of the process, not only the end. Analysts define the 44 ESG sub-risks, design how severity is assessed, and fine-tune the models. The methodology is a human construction that machines then apply consistently at scale.

A Division of Labor, Not a Contest

Seen this way, the old debate dissolves. The question was never whether AI or people should produce ESG ratings. The question is which part of the work belongs to which.

Machines provide reach, consistency and speed. They apply the same rules to every entity, every day, without fatigue or favor.

People provide judgment, correction and improvement. They decide what the system should look for, they catch what it gets wrong, and they raise the standard of the model over time.

The result is a rating that is both broad enough to cover the real world and rigorous enough to be relied upon. Scale without rigor is noise. Rigor without scale never reaches most of the companies an investor actually holds. The value is in the combination.

What This Means Going Forward

The EU ESG Rating Regulation asks providers to disclose how their ratings are built. SESAMm has chosen to disclose the full pipeline, including the role of AI, the points where it can fail, and the human controls that contain it. The aim is not to claim the technology is flawless. It is to show, in detail, why the output can be trusted anyway.

As artificial intelligence becomes more capable, the temptation to remove the human layer will grow. SESAMm's position is the opposite. The more powerful the models become, the more valuable the people who direct them, check them and teach them become. That is the architecture of a rating worth trusting, and it is now open for anyone to read.

To explore the full methodology behind the Controversy Exposure Score, visit sesamm.com/methodology.