India to Launch Four Responsible AI Solutions on AIKosh Platform

India to Launch 4 Responsible AI Solutions on AIKosh From September | AIM

These selected applications will span critical sectors such as healthcare, agriculture, climate change, governance, and support for learning disabiliti In a significant step towards ethical artificial intelligence development, the Indian government will launch four responsible artificial intelligence (RAI) solutions between September and December this year under its IndiaAI Mission's 'Safe and Trusted AI' pillar. These tools will be made available through the government's dedicated AI platform, AIKosh, focusing on machine learning, bias mitigation, risk evaluation, and fairness assessment, IndiaAI Mission chief operating officer Kavita Bhatia told ET. She also noted that the government is currently evaluating over 400 applications for AI tools, including those for deepfake detection and watermarking AI-generated content. Once fully developed, these tools are intended to be released as open-source resources. Alongside the rollout, the government will also select 30 AI application development projects within the next month as part of a broader initiative to promote real-world use cases, the official reportedly said. These selected applications will span critical sectors such as healthcare, agriculture, climate change, governance, and support for learning disabilities. The Safe and Trusted AI initiative is one of seven key pillars of the ₹10,000 crore IndiaAI Mission. Other components include the IndiaAI Innovative Centre, Datasets Platform, Futureskills Programme, Startup Financing, and Compute Infrastructure, all of which aim to build a robust national AI ecosystem. Moreover, AI experts and developers have praised India's leadership in responsible AI. At a recent discussion with AIM, Alpan Raval, chief AI/ML scientist at Wadhwani AI, shared how an oral reading fluency assessment tool developed using AI4Bharat's open-source models has already been deployed in Gujarat's public schools. "India has taken the lead in using AI for social good -- not just in theory but in actual implementation," Raval remarked. Sourav Banerjee, CTO and co-founder of United We Care, echoed this sentiment, noting that ethical AI practices are becoming more integral to the development process across sectors. To support this vision, Infosys and the Linux Foundation have also partnered to promote India's homegrown Responsible AI solutions on the global stage. As the AIKosh portal prepares to go live with its first batch of tools, the IndiaAI Mission continues to establish India's position as a frontrunner in the global push for ethical, inclusive, and secure AI.

Four of IndiaAI's responsible AI solutions to go live on AIKosha from September

Safe & Trusted AI and IndiaAI Application Development Initiative are among the seven pillars of the Rs 10,000 crore IndiaAI Mission. The others are the IndiaAI Compute Capacity, IndiaAI Innovation Centre, IndiaAI Datasets Platform, IndiaAI FutureSkills and IndiaAI Startup Financing. The government had last year invited expressions of interest for RAI tools and received about 2,000 applications. Four of the eight responsible artificial intelligence (RAI) solutions developed under the IndiaAI Mission's safe and trusted AI pillar are expected to go live on the mission's AIKosha portal between September and December. The government is also aiming to make available the 30 projects selected under the application development initiative within a month as "try models", said an official. The RAI solutions to be made available are in areas such as machine unlearning, bias mitigation, risk assessment and fairness assessment, IndiaAI Mission chief operating officer Kavita Bhatia told ET on the sidelines of a consultation on AI readiness assessment methodologies in Delhi on Tuesday. Deepfake detection tools are still in the works, according to her. More than 400 applications have been received for areas like deepfake detection tools and watermarking of AI-generated content. These are under evaluation and the aim is to make all such tools open-source, she said. The 30 selected AI applications under the application development initiative are under the themes of healthcare, agriculture, climate change management, governance and learning disabilities. Safe & Trusted AI and IndiaAI Application Development Initiative are among the seven pillars of the Rs 10,000 crore IndiaAI Mission. The others are the IndiaAI Compute Capacity, IndiaAI Innovation Centre, IndiaAI Datasets Platform, IndiaAI FutureSkills and IndiaAI Startup Financing. The government had last year invited expressions of interest for RAI tools and received about 2,000 applications. The eight selected participants include IIT Jodhpur, IIT Roorkee and IIT Delhi. More than 900 submissions were received for the application development initiative, according to the electronics and information technology ministry's statements. "We need to assess what is the readiness with regard to making AI applications that are inclusive, that are responsible, that are safe, that are trustworthy and which follow the ethical framework," said Abhishek Singh, CEO, IndiaAI Mission, speaking at the consultation. While support for AI models, computing and language application programming interfaces are there, the solutions need to be fair and meet real needs on the ground, he said, reiterating that India's approach to AI regulation is pro-innovation and light-touch with a focus on preventing harm. The government is collaborating with UNESCO and Ikigai Law to produce a report on an AI readiness methodology for India. The final stakeholder consultation for this was held in Delhi on Tuesday. Consultations had earlier been held in Bengaluru, Hyderabad and Guwahati. "Globally, we know that AI systems are being used to allocate welfare benefits, screen job applications, generate educational content and even inform judicial and healthcare decisions," said Tim Curtis, director and UNESCO representative for the UNESCO regional office in Delhi. "In India's context, if these tools are not designed with a whole spectrum of users in mind, including marginalised communities, they really do risk leaving a lot of people behind." The risks could undermine AI's immense benefits if proactive measures are not taken for responsible development and usage, he added.

IndiaAI Mission: Four Responsible AI Solutions To Go Live From September

The government is also aiming to roll out the 30 projects selected under the application development initiative for trial phase Of the eight responsible artificial intelligence (RAI) solutions built under the IndiaAI Mission's 'safe and trusted AI pillar', four solutions are expected to go live on the mission's AIKosha portal between September and December. An ET report, citing IndiaAI Mission's chief operating officer Kavita Bhatia said that these four solutions to be made available are in the areas, including machine learning, bias mitigation, risk and fairness assessment. She further added that deepseek detection tools are still in the making. Besides, the government is also aiming to roll out 30 projects selected under the application development initiative within a month as "try models". These projects are under the theme of healthcare, agriculture, climate change management, governance and learning disabilities. For the deepseek detection tools, Bhatia added that the government has received more than 400 applications, which are under evaluation and likely to be open-source. Last year, the Centre selected eight strategic projects to promote responsible development, deployment and adoption of AI in the country. Back then, the government received 2,000 proposals in response to its expression of interest (EoI). These projects aim to address the challenges of bias in AI, enhance privacy, and establish frameworks for explainable and trustworthy AI. Under the IndiaAI Mission, the Centre has selected four AI startups - Sarvam AI, Soket AI Labs, Gnani.ai and Gan.ai, to build large language models (LLMs). While Sarvam AI was the first startup to be picked by the government, the other three made it to the list last week. Additionally, the government has shortlisted seven companies for the second round of GPU tender under the mission. These companies, including Sify Digital Services, Netmagic IT Services, Vensysco Technologies, Cyfuture India , Yotta Data Services, Locuz Enterprise Solutions, and Ishan Infotech, cleared the technical presentation round. IndiaAI Mission - A Look Back In 2024, the government launched IndiaAI Mission with an initial outlay of INR 10,372 Cr over the next five years. With this mission, the Centre aims to build an AI ecosystem, offering supercomputing capabilities comprising over 10,000 GPUs to various stakeholders including startups, academic institutions, government agencies and research institutes. In the first bidding round, the Ministry of Electronics and Information Technology invited bids from cloud services providers and data centre operators to supply and manage AI computing infrastructure including GPUs. The Centre selected ten tech giants, including Jio Platforms, Tata Communications, CMS Computers India Pvt Ltd, Ctrls Datacenters Ltd, E2E Networks Limited, Locuz Enterprise Solutions Ltd, Orient Technologies Ltd, NxtGen Datacenter and Cloud Technologies Private Ltd, Vensysco Technologies Limited and Yotta Data Services Limited. Last month, the government said that it has received 506 proposals to build the AI foundational models across three phases.