History has shown, time and time again, how humans and machines have coexisted, and how humans have leveraged the power of machines to improve their quality of life. From water-lifting devices created by ancient civilizations to access and distribute water to the Industrial Revolution where rapid developments in machinery increased production at scale by automating manual tasks and minimizing human effort. These interwoven relations and traditions between humans and machines have been ingrained in our societies for thousands of years.
AI will not Replace Human Intelligence
Fast forward to the 21st century, where AI proliferates the world we live in. The natural existential question we ask is, “Will AI replace humans?” Are humans, in fact, competing with machines? If we look back at history, the clear answer that emerges is “no”.
Lessons from our ancestors suggest that humans have always invented ways to outsource our physical, manual tasks to machines in order to increase efficiencies at scale. In this AI-driven world, we now simply outsource our intelligence to machines to automate the tedious, mental tasks. The tenor underlying the question of whether AI will replace humans, however, suggests that there is an equivalency between humans and machines, and this is arguably questionable. For instance, AI-assisted machines have more computing power, can execute computations faster than the human brain, and can complete processing tasks in a more consistent and accurate manner. Conversely, humans have emotions, intuition, and can understand nuance, ambiguity, cultural and societal sensitivities. Intelligent machines cannot by themselves develop these attributes, and with the help of humans providing some form of ‘training’, may be able to mimic some of these attributes. Nevertheless, working together, humans and machines can radically accelerate progress in leaps and bounds, as history has systematically demonstrated time and again.
AI in ESG and Degrees of Automation
The need for accelerated progress has never been more critical than now. Whilst companies, investors, governments, and stakeholders around the world scramble to solve the climate, environmental and human crises, it’s time to, once again, explore how humans can leverage machines to help address these problems.
The lack of a global standard framework for ESG disclosure and reporting has provided some hurdles to prevent the collection of data in a consistent and timely fashion to drive informed decision-making. The establishment of the International Sustainability Standards Board (ISSB) and the new EU ESG disclosure mandates like Sustainable Finance Disclosure Regulation (SFDR) and Corporate Sustainability Reporting Directive (CSRD) catapult ESG disclosures to the forefront.
But, how can the ESG disclosures be searched efficiently, and the related quantitative and qualitative data be extracted and structured to drive decisions?
Meet idaciti ESG Accelerator
At idaciti, we have developed a ‘human-in-the-loop’ machine-learning platform (i.e., ESG Accelerator ™). Our platform simulates the search workflow that we are all accustomed to when searching for relevant information on the web:
(1) we type our search phrases, (2) the search engine retrieves relevant search results in seconds; and (3) we examine some of the returned results on the first or second page to find the most relevant content.
Our ESG Accelerator does just that, and more - the user types in a search phrase or phrases, and the search engine scours millions of documents, returning only the relevant results in seconds. The user identifies some good or relevant examples of disclosures and provides these good examples back to the platform for it to learn. In essence, the rich domain knowledge and expertise of the user are now embedded into the platform to make it more ‘intelligent’.
The platform can now look for similar examples of these disclosures within seconds, and extract structure and make the qualitative and quantitative data available for further analyses by the user. In addition, leveraging advanced NLP algorithms, the platform generates a confidence score for the extracted data so that users can assess its accuracy and know that they can rely on it for analyses and decision-making.
This overall workflow is not all that different from the historical outsourcing of manual, tedious tasks to the machine. Now, we just use fancier nomenclature and acronyms to discuss and describe it.