Responsible AI cannot be an afterthought, needs to be part of DNA: Infosys CTO Rafee Tarafdar | Technology News

Mohammed Rafee Tarafdar is CTO of Infosys, and is responsible for the technology vision and strategy at the company. He founded the Strategic Technology Group, a global team of technology leaders, full stack architects and power programmers working on complex digital, cloud and AI programmes for enterprise clients of Infosys and Infosys platforms.A Mechanical Engineering graduate from Karnatak University, Rafee Tarafdar is passionate about social impact and personally mentors the Tech For Good programme at Infosys.
He spoke to on how Tech for Good fits into the organisation’s goals and objectives, on Responsible AI, and how AI could impact the common man. Edited excerpts:
Story continues below this ad

Venkatesh Kannaiah: Where and how does Tech For Good fit into your organisation’s work and goals?
Mohammed Rafee Tarafdar: Tech for Good falls under Infosys’ ESG ambitions and goals. When we formulated and published our ESG vision for 2030, there were multiple pillars and one of the pillars is using technology to drive social good.
As per our plans for Tech for Good for 2025, we had a goal to impact around 80 million lives. We have achieved our goal. Now, we are trying to redefine our goals for 2030 and will publish it in the next few months.
Tech for Good, for us, is a strategic area of focus and we believe that, given all of the tech that we work with clients globally and in India, we can use it to drive social impact and change. And for this, we have identified three broad areas for impact: e-governance, healthcare, and education. We have, on average, 50 dedicated professionals who work solely on Tech for Good initiatives. We also bring in volunteers from other teams for specific projects. I personally lead the initiative at an organisational level.Story continues below this ad
Venkatesh Kannaiah: Can you give us some instances where your Tech for Good work with governments or organisations has helped improve public services?
Mohammed Rafee Tarafdar: We develop and open source the software that will impact larger societal goals. On the e-governance front, we are partnering with Indian Institute of Information Technology – IIIT Bangalore, where we work on their Modular Open Source Identity Platform (MOSIP) which is essentially building a digital identity for citizens across the world. We are working with them on three specific areas.
We have built an Android regration client, which can be used to onboard users on to their digital identity platform. And this, today, is being used in places like Zambia and Ethiopia, solving the problem of onboarding customers or users in remote areas with poor internet connectivity. So this is an integrated hardware-cum-software kit. This is what the agents in different countries will take and go to their remote villages and onboard citizens onto the respective digital identity platforms of these countries.
Secondly, we are working with MOSIP on a module for enabling single sign-on for using the respective country’s digital identity platform. For example, if you want to avail any government service in these countries, can you use this to authenticate yourself or to say who you are and the system would immediately identify you and you would be eligible for the services.Story continues below this ad
Third, we’re also working with MOSIP for certifying, verifying, and managing digital credentials; for example, it could be used to verify a farmer’s certificates and details of his farm. The way we do this is, we generally identify modules that our Infosys Tech for Good team can work on. We work with the MOSIP team, write the code in open source, which they then take it and integrate it into their platform and dribute it to different countries/stakeholders building digital identity initiatives across the world.
Venkatesh Kannaiah: Can you tell us more about your work in the health domain and in assive tech in particular?
Mohammed Rafee Tarafdar: We have partnered with the LV Prasad Eye Institute and have built an app called SiteConnect, using which any person can check their eyes for vision, cataract, red eye, or other issues. It will help identify the issue and then direct the person to a visual care centre. The idea is to democratise the tech and make it accessible. The ophthalmology expertise is coming from LV Prasad Eye Institute. Around 1. 5 lakh users across the country have used it and almost 40 percent of them have moved for further medical intervention.
Venkatesh Kannaiah: Can you tell us about your work in the sphere of education?Story continues below this ad
Mohammed Rafee Tarafdar: We have a programme called Springboard, an open source platform for digital skilling of citizens. We have made it contextual and localised it for different Indian languages. Our goal was to reach around 10 million citizens and help reskill them. We have on the ground activities too with Springboard. Our team is partnering with and working in Chamarajanagar drict in Karnataka to skill prison inmates. We also have Springboard Lab on Wheels, a mobile unit where we use Raspberry Pi devices in order to enable education in villages.
Venkatesh Kannaiah: Tell us about the Tech for Good innovations and products that have been showcased your Aarohan awards?
Mohammed Rafee Tarafdar: Aarohan is a social innovation award that we do as part of Infosys Foundation. The goal for us is to identify innovators who are using technology to create an impact. We identify these innovators and provide them with tech and monetary support. In 2023, we identified eight such innovations and we are supporting them.
Venkatesh Kannaiah: What are your thoughts on Responsible AI? What could be the roadmap?Story continues below this ad
Mohammed Rafee Tarafdar: We started the AI-first initiative at Infosys and we were very clear that we want to be responsible design. Everything that we build within Infosys, be it an AI model, AI product, or AI agent, we want to follow Responsible AI protocols. There are 12 areas that we focus on from a Responsible AI standpoint. We want to be doing it right and also be compliant. We set up a Responsible AI office and we have a head of Responsible AI who ensures that what we do at Infosys is right according to these 12 parameters. Responsible AI cannot be an afterthought, it has to be a part of the DNA.
We found out that a lot of these AI models have vulnerabilities, and have built technical guardrails. We have now open sourced a lot of these guardrails for others to learn from and use it accordingly.
Venkatesh Kannaiah: Can you give your outlook/prediction on three areas of a common man’s life in India which will change through AI?
Mohammed Rafee Tarafdar: In India, we have a lot of regional languages and one thing that is likely to change is how we plan to provide access to citizen services in regional languages or the language of choice. So these citizen services could be provided in regional languages voice-based interactive systems. Suppose there is a visual impairment, the question is whether I can perform these tasks through voice? Can some software read out the information for me and can I interact with the system and get what I want from it? This is going to be a big leap using AI.Story continues below this ad
The second is on how we could use AI to improve the quality of basic services, whether it is in providing water, electricity, or sanitation. And in a lot of these, I think AI can help us with better prediction, better decision-making processes, better data analysis to improve the quality of life in meaningful ways. Because a large part of our population still lives in villages and towns, I think we need to focus on frugal innovations. It is all about how we create these small AI embedded devices, which can then be used in the agriculture sector to improve crop yields or enhance milk production. Can we use this to improve weather prediction or manage floods better? I think having these frugal AI innovations at the edge will make a huge difference.
The question that then arises is how long will it take for such an impact? Let us take the issue of e-governance in local languages. One issue is to do with the breadth of the coverage. Today, for instance, the question is whether all languages and all dialects are digitised. If some parts of it are not digitised, it may not work well. Second is the accuracy levels we are able to reach within each language, which also needs to improve. What I have seen over the past two years is that this tech is improving very fast. We are seeing improvements every few weeks.
Venkatesh Kannaiah: What are the three technologies you would bet on to see substantial social impact in an Indian context?
Mohammed Rafee Tarafdar: AI is obviously the first because this helps us make these citizen services accessible in vernacular languages and in the natural way of speaking and communicating.Story continues below this ad
Second is edge computing. We see that while urban areas are well connected, there are still a lot of remote villages and towns. I think we need to have the ability to run a lot of technology at the edge, to drive the accessibility of technology till we resolve connectivity issues.
The third is robotics. I think today, with the level of innovation that is happening in robotics as well as in humanoids, if we can frugally implement robotic solutions, it will be a gamechanger.
These autonomous machine robots could be used to automate tasks, and bringing cognitive and robotic aspects together, we can change a lot of things in farming, cattle management and milk production.
I know that some startups are already working on using a combination of IoT and AI in order to improve the yield. But, with AI, it takes some time to get to a significant level of maturity and accuracy. So we should continue on this path and keep investing and training for it to become more meaningful.
Venkatesh Kannaiah: On driverless cars?Story continues below this ad
Mohammed Rafee Tarafdar: From our experience, what we have found is that such things work better in controlled environments like a factory, airport or a warehouse. We are already finding that they can add significant value in such locations. At Infosys, we have about 50 autonomous machine robots that are being used in some of these controlled environments. So I think that is where we’ll immediately see the value.