The ever-evolving artificial intelligence (AI) landscape has generated excitement, interest and investment; it has also triggered important questions regarding AI’s impact on businesses, on investment portfolios, on society and on the environment. When factoring in elements such as the high energy requirements of training large language models (LLMs), direct and indirect emissions and the diversity of typical AI company boardrooms, the intersection of AI technology with environmental and governance concerns highlights a complicated ethical landscape. Access to robust datasets becomes critical to measure and assess the likely answers to these questions.

Indices can play a major role in revealing deeper insights about specific industries or investment themes such as AI. Offering a comprehensive perspective on companies involved in the space, the S&P Kensho Global Artificial Intelligence Enablers Index (S&P Kensho Global AI Enablers), was launched in October 2023 and currently comprises 37 companies at the forefront of developing and enabling AI technologies (see Exhibit 1).

(Click on image to enlarge)

In part due to rising demand for AI-optimized servers, voice technologies, data centers and more, the S&P Kensho Global AI Enablers Index has outperformed the S&P Global BMI Information Technology (Sector) and the S&P 500^® since its launch (see Exhibit 2).

(Click on image to enlarge)

As the spotlight brightens on AI’s potential, questions about sustainability have also come under scrutiny. While ESG metrics do not equate to ethics, they do offer valuable insights into the consequences of AI companies’ business practices. As measured using S&P Global ESG Scores, overall, the S&P Kensho Global AI Enablers Index currently holds lower “E,” “S” and “G” scores than the S&P Global BMI Information Technology (Sector), S&P TMI Information Technology and S&P 500 Information Technology (see Exhibit 3).

(Click on image to enlarge)

Looking more closely at the Governance score, the S&P Kensho Global AI Enablers Index scored poorly on the Board Gender Diversity metric, with a weighted average of only 30% women on AI constituent boards. In some cases, there are no women at all. This index’s average is even lower than that of the notoriously male-dominated U.S. tech sector. However, the lower average doesn’t tell the whole story. In fact, gender diversity scores in the AI index are highly dispersed and highlight that there are a number of companies that do relatively well on this metric compared to other indices (see Exhibit 4).

(Click on image to enlarge)

Many AI companies also differ from their benchmarks on environmental metrics. Notably, training AI may not burn fossil fuels directly, but it can take a lot of computing power (and hence energy), potentially via a third party (e.g., cloud computing services). Along with market convention, S&P Global Trucost’s carbon data set breaks down corporate emissions into Scope 1, Scope 2 and Scope 3. Scope 1 includes direct carbon emissions from owned sources; Scope 2 includes indirect emissions from purchased electricity; and Scope 3 emissions involve all other indirect emissions in the value chain that can be divided into upstream and downstream. Indirect emissions classified as Scope 3 upstream occur from supply chain activities such as procurement and logistics, and Scope 3 downstream emissions stem from product use (see Exhibit 5).

(Click on image to enlarge)

In addition to the emissions data, S&P Global also provides data on temperature alignment. While carbon intensity measures current greenhouse gas emissions relative to economic value, the Temperature Alignment metric assesses how well a company’s emissions trajectories align with global temperature targets, such as 1.5°C or 2°C, thus focusing on long-term climate goals. This metric is key in assessing a company’s commitment and capability to mitigate climate risks, as well as its alignment with global climate goals. The S&P Kensho Global Artificial Intelligence Enablers Index is aligned with 1.5°C scenario, which demonstrates the lowest temperature rise (see Exhibit 6) and a more favorable position than the global tech sector index.¹

(Click on image to enlarge)

Ultimately, the ethics and environmental impact of AI companies is nuanced, reflecting both extensive energy requirements and efforts to improve efficiency and sustainability, as well as the relative efficiency of AI-linked revenues. Indices such as the S&P Kensho Global Artificial Intelligence Enablers Index, and the data perspectives powered by S&P Global ESG Scores and climate data sets, can help market participants to not only understand how AI drives market performance, but also to assess its potential impact in a broader context.

¹ For more on carbon intensity and temperature alignment, see Maya Beyhan’s blog, Schrödinger’s Carbon: Intensity and Paris Alignment.

More By This Author:

The S&P/ASX Index For “Larger,” More Liquid Small Caps
Time, Trust And Trading
S&P Momentum Indices Leading The Pack