Four Dimensions of Responsible AI

Artificial intelligence (AI), once largely a niche interest of computer scientists, is now everywhere. Although AI is hardly new, the recent convergence of immense data availability with unprecedented computing power has led to an explosion of AI use cases. As advances in data storage and transfer technologies continue to accelerate, the sheer amount of data being processed by AI systems will only skyrocket.

Recent advancements in user-friendly generative AI have led to the near-instantaneous production of computer-generated text, images, and audio. This emerging wave of AI innovation has sparked a frenzy of new applications and developments, with some products gaining widespread adoption at warp speed. The inertia of this AI revolution is unstoppable, and AI is already delivering transformational benefits and efficiencies to individuals, companies, and governments.

At the same time, AI tools are raising numerous ethical and legal questions. No part of society — from democratic norms and labor markets to educational institutions and scientific progress — will remain untouched. Yet AI brings both promises and risks. To harness the former, we must mitigate the latter. And we must do so using a Responsible AI framework.

For Wipro, Responsible AI encompasses four dimensions:

1. Individual
2. Social
3. Technical
4. Environmental

1. Individual Dimension

AI tools must protect and enhance human dignity.  They must safeguard privacy and champion the security of personal information. They must not discriminate or undermine an individual’s identity, language, or background.

Protecting the dignity of the individual means that we must be open and transparent about our AI tools, clearly communicating how they use data and why our due diligence processes ensure trustworthiness. Such practices do more than facilitate legal compliance — they also enable citizens to exercise proper scrutiny of AI processes and outcomes, which is essential to creating an environment of trust.

Our approach embraces:

• Privacy and security by design, emphasizing privacy-enhancing technologies
• Fairness in processing and in the outputs of AI tools, achieved through rigorous assessments and testing
• Responsible data handling: preparing the right datasets with equality and equity in mind
• Rigorous safeguarding of copyrighted information
• Vocal promotion of responsible and critical use

2. Social Dimension

AI systems must improve society and strengthen the values that bind us together.

Our approach embraces: 

• AI impact assessments that bring together data, systems, and people, starting with the people and communities that will be impacted the most
• AI for good: a strong focus on tools that tackle the world greatest challenges and help companies increase productivity
• Transparent AI that creates an ecosystem of trust in which people understand how AI systems are made and where the data originates
• Strong support for measures that avoid the proliferation of misinformation, disinformation, and harmful deepfakes
• Involvement of communities and employees in considering the economic and labor force impacts of AI

3. Technical Dimension

AI systems must be robust and safe, and they must protect both personal and company data.

Our approach embraces:  

• Security by design, so that security is embedded in every AI project from the start
• Strong protection against attacks that could undermine privacy, pollute outcomes, or lead to unfairness and discrimination
• Strong governance through a cross-organizational approach to ensure that AI tools and systems are properly handled, governed, and monitored
• Compliance with laws and policies applying to AI, including those related to privacy, human rights, liability, consumer legislation, and working to remain compliant as new guidelines and laws being developed.

4. Environmental Dimension

Achieving environmentally sustainable AI will require compute-intensive AI systems to limit their footprint on the environment.

At Wipro, we champion:    

• The use of the cloud computing as a more energy efficient option
• Data minimization and smart data processing approaches to reduce the environmental impact
• A push for synthetic data to limit data extraction
• A focus on AI systems that help us tackle climate change issues, the greatest environmental challenges of our time