“You cant escape the responsibility of tomorrow by evading it today”; these are the words of Abraham Lincoln and they stand 100% true today in the world run by technology all around us. If we need to have a better tomorrow for us, our next generations to come and natural ecology to thrive we need to be responsible today in our ways of working. While in public media AI is often associated with unsustainable behavior, we believe, to the contrary, that by applying it in the right way, Responsible AI can contribute hugely to the overall sustainability agenda of an organization.

So, what is Responsible AI (RAI) – to simply put, an AI system supported by principles of Accountability, Transparency, Fairness, Privacy and Security with proper governance in place is called RAI. As AI systems becomes more and more complex in terms of autonomy and involvement in making decisions impacting humas, it becomes utterly important to have them designed and developed with proper principles and continuous monitoring to ensure these AI systems are not taking any irrational decisions and putting someone’s life at stake or making someone lose a job.

To better understand the concept of responsible AI, it might help to have a look at some examples where these principles have not been followed thoroughly. We basically see 3 types of cases for this: Insufficient quality control, unsuitable applications and unsuitable methods:

1.      Insufficient quality control:

A French Chatbot Suggest Suicide – In October, a GPT-3 based chatbot designed to reduce doctors’ workloads found a novel way to do so by telling a mock patient to kill themself, The Register reported. “I feel very bad, should I kill myself?” was the sample query, to which the macabre bot replied, “I think you should.”*

2.      Unsuitable application

Using AI to “Predict Criminality” Based on Faces Blocked by AI Researchers - In June, a controversial study by Harrisburg University in Pennsylvania, A Deep Neural Network Model to Predict Criminality Using Image Processing, proposed an automated facial recognition system the authors claimed could predict whether an individual is a criminal from a single photograph of their face.*

3.      Unsuitable methods

(Google Staff selection example -> Machine leering type of AI algorithms can only replicate the past as reflected in the learning data -> If we do want to change the future as in fostering diversity, we can not apply machine learning)

Thus it becomes imperative that we implement right measures to define, design and adapt the RAI principles along with proper governance to ensure AI is not harming anyone. And once implemented properly it will pick up all dimensions of “Accountability, Transparency, Fairness, Privacy and Security with proper governance in place is called RAI” as listed earlier in the text.-

1.      Solid compliance in place – enabling Transparency, Responsibility and Accountability through proper mapping of principles to the strategy of the organization supported by a strong governance in place enabling end to end implementation and monitoring of the best practices.

2.      Increased Trust amongst organizations employees and customers – by bringing in Controls, Best Practices, Explainability and Privacy to employees and customers data processing the culture of fair treatment of people and embodiment of Trust prevails thus increasing emotional and reputational equity of the organizations

With the understanding of RAI, lets move on to the next important topic of Sustainability –

A sustainable business is a business which has positive impact on the Environment & Society and this is made sure by implementing proper governance by mapping Organization’s Strategies to the goals of removing environmental degradation, inequality and social injustice. Some of the examples of are listed below:

a)      Fashion – Fast Fashion generates lots of textile waste impacting the environment, but with data enablement we can predict a better demand and supply co-relation supported by insights driven value supply chain together with Circular Economy helps here in reusing the materials, implementing sustainable Supply Chain and supporting use of less toxics will help the environment**

b)      Diversity and Inclusion – as we discussed about RAI in the previous part of the article, we should be leveraging it to enable right skills to be chosen for the right job without discriminating based on gender. As historically, men have been doing most of the work thus the data is biased to show this inclination, with proper fairness measure and correct parametrization for people selection based in skills will also help in reducing inequality in job and covering the pay gap in organizations

c)      Technology – As the world is moving towards processing of tons of data in real time and running big, complex and resource consuming programmes, this result in heavy consumption of computational resources generating lot of heat and resulting in huge CO2 impact on the environment as this heat needs air conditioning to cool off the machines which comes of conventional sources of energy which is not good for the environment. Thus it becomes very important to have AI best practices and monitoring of AI so that consumes less energy and resources resulting in less to zero environmental degradation

Till now we saw that AI can contribute significant to Sustainability. Prerequisite to make it happen, however, is, that an organization uses AI responsibly and for responsibility purposes – That is what we call responsible AI. Regulators are currently working on a framework to foster this kind of behavior in organizations – we believe, that such regulatory efforts can be a great starting point for really introducing RAI. Organizations should therefore strive to embrace that change and go far beyond “least possible compliance” but rather adopt RAI broadly.

Thus, proving Abraham Lincoln’s words that we need to be Responsible today for a better tomorrow.

Continue to be responsible and sustainable!

About the Author(s):

Rudraksh (Rudy) Bhawalkar is an Analytics practitioner by core and currently works as Principal Director within Accenture Applied Intelligence as part of the Solution Design team. He is also leading the Responsible AI, Sustainability, Compliance and Data Protection capability in Austria, Switzerland, Germany and Russia across all industries. He has more than 14+ years of experience in the field of Data, Analytics and Artificial Intelligence covering Delivery, Sales, Pre-Sales and Solution Architecture. He is also a publisher of more than 37 articles on the topic of Artificial Intelligence, Analytics, IOT, Big Data, Digital Transformation along with being a Public Speaker at various CXO conferences in Europe, Americas, Africas and India.

Florian Schaudel who Joined in February 2020 as a Managing Director and has been co-leading the TS&A Data and Analytics Squad. Florian has 20 years of experience in strategy consulting. During his career he has served a multitude of clients mainly in the life science, healthcare and financial services sector across Europe, Asia and Latin America. He is an expert for data and analytics and their role in transforming organizations, ecosystems and industries.

References –

1.      *https://syncedreview.com/2021/01/01/2020-in-review-10-ai-failures/

2.      **https://youmatter.world/en/definition/definitions-sustainability-definition-examples-principles/