Action Plan: Building a Diverse AI and Data Science Workforce

The AI and Data Science higher education landscape in India is changing rapidly. While a number of India’s higher education institutions are well on their way to developing new AI and DS programmes, other institutions are just beginning the journey [7]. With the gap between supply and demand of individuals with AI and DS knowledge and skills expected to grow significantly, it is imperative that a wider set of stakeholders including higher education institutions, professional societies, industry, start-ups, and government departments come together to define the common contours of AI and DS curricula that can serve as the foundation for innovative programmes across a large number of Indian institutions. Academic institutions in the United States are further along the road on this journey, and Indian institutions can benefit from their experiences, leveraging best practices in education and training.

Programs that prepare professionals in cutting edge disciplines at the interface of big data and machine learning and domains such as health, environment, and sustainable living are a priority under India’s NEP 2020 [8]. This is an important policy statement as it highlights the need for interdisciplinary programs and integrated curricula that focus on the implementation of AI and DS tools in specific application domains. The U.S. has several successful models including the University of California, Berkely’s Division of Computing, Data Science, and Society whose “dynamic structure connects computing, statistics, the humanities, and social and natural sciences to create a vibrant and collaborative environment that accelerates breakthrough research across scientific and technological frontiers.”

U.S. and European AI and DS models can serve as a framework for adoption and implementation by Indian institutions. However, understanding the unique challenges in the Indian context and identifying the knowledge and skills gap will be critical to developing innovative programs to address India’s workforce priorities.

The IUSSTF/ itihaasa exploratory survey provides an overview of the AI and Data Science higher education landscape in India and identifies some of the gaps and challenges. While the geographic diversity of institutions that responded to our survey was encouraging, one cause for concern was the limited number of interdisciplinary programs in the priority areas of health and agriculture.

To address some of the challenges and gaps and identify potential solutions, we propose a series of action items that will engage the broader stakeholder community and lead to more substantive conversations about the AI Workforce.

Action Item 1

Conduct a visioning workshop on “Developing a Diverse, Robust AI Workforce: Innovations in AI Training and Program Development”.

One of the goals of IUSSTF’s USIAI initiative is to build on key priorities, recommendations, and common challenges articulated in the AI Strategic Plans of the two countries. USIAI provides a platform to engage key stakeholders, including government agencies, academic institutions, industries that are creating/adopting AI tools and technologies, think tanks, professional societies, and foundations.

IUSSTF and itihaasa are planning an Indo-U.S. Visioning Workshop that will bring together faculty from Indian and U.S. academic institutions and representatives from industry to address the challenges associated with developing a diverse, robust workforce. The workshop will focus on specific aspects of training and program development under four broad themes:

» AI, Data Science training for Industry

» Undergraduate Level Programs

» Graduate Level Programs – Computer Science focused

» Graduate Level Programs – Other Engineering / Science / Research focused

The goal of the proposed workshop is to create a roadmap for developing a diverse, globally engaged, AI workforce through the following objectives:

» Identify the knowledge and skills required for different AI and DS careers and recommend different education and training pathways.

» Map domain-specific and sector-specific AI and DS knowledge and skills

» Share best practices between India and the U.S. for teaching and training

» Build a more inclusive and diverse AI/DS discipline

» Recognize the key challenges in AI and DS education and training, and develop strategies to mitigate these challenges

» Identify synergies for collaborative research and training activities

Last, but not the least, we hope the workshop will foster individual and institutional partnerships and create a vibrant community passionate about education and training.

Action Item 2

Study the AI and DS programs in India and the U.S.: Identify exemplars - course curriculum, datasets and tools, resources

Chapter 2 provided a detailed summary of U.S. and European initiatives addressing competencies and curricular guidelines for AI and Data Science. Our survey provides some preliminary insights into the structure of AI and DS programs in India. We recommend a detailed follow-up study to identify building blocks of Bachelor’s and Master’s programs in AI and DS on the basis of the course content and learning outcomes from a representative sample of universities in India. This study will also compare and contrast the curriculum at U.S. and Indian universities and identify significant gaps/ challenges, if any, in the Indian context. The study will serve as a resource for accrediting bodies and industry groups and help in the development of new India-specific guidelines for curriculum and training.

Action Item 3

Organize follow-up workshops/meetings on AI and DS Workforce Development

Ramping up India’s AI and DS workforce will require the sustained efforts of different stakeholders including the Department of Science and Technology, Ministry of Education, All India Council for Technical Education (AICTE), Ministry of Electronics and IT (MeitY), academic institutions, professional societies, and industry.

Industry, the largest employer of AI and DS talent in India, must play a key role in the development of new programs as the knowledge, skills, and competencies vary across different sectors. Collaboration with industry was identified as the third most important factor impacting AI and DS higher education programs in our survey. This is a challenge faced by many U.S. institutions as well. Organizations such as the U.S. India Business Council, U.S. India Strategic Partnership Forum, NASSCOM, and FICCI can identify successful models and mechanisms for industry-academia collaborations.

Bilateral workshops on critical issues such as Explainable AI, Ethics and Trust, Guidelines for data sharing and benchmarking can help in the development of a shared framework for enabling impactful and responsible innovation in AI.

Action Item 4

Accelerate efforts to scale-up computing infrastructure and access to large datasets.

The top two challenges impacting AI and DS programs in India in our survey are (1) access to high end computing infrastructure and software, and (2) availability of quality datasets. This is consistent with the findings from itihaasa’s study on the landscape of AI/ML research in India [11].

Setting up world-class computing infrastructure for AI/ML research is expensive. India’s MeitY and NITI Aayog have programs to ramp up shared computing infrastructure [22] [23]. In recent years, a total compute capacity of 22-petaflop spread across 15 educational and research institutions, including a National AI facility at CDAC Pune, have been installed. Another 32-petaflop spread will be deployed with indigenously designed systems. The NITI Aayog report recommended a national computing infrastructure, AI Research, Analytics and knowledge Assimilation (AIRAWAT), a 100-petaflop supercomputing system for AI applications [9].

While the Indian computing infrastructure is ramping up, U.S. cloud computing infrastructure companies could fill the gap by offering affordable plans to Indian higher education institutions under a special agreement between stakeholders on both sides. India is already a large market of many of these companies and an investment in the future AI workforce would certainly be mutually beneficial.

Collaborations between India and the U.S. can accelerate the development of data infrastructure and resources to facilitate AI and DS research and training. While publicly funded AI/DS research in India may focus on India-specific challenges, some of the solutions may be relevant in the U.S. context, especially in areas such as health equity, energy, and climate. India’s draft National Data Governance Policy formulated by MeitY proposes the launch of a non-personal data-based India Datasets program and addresses rules to ensure that non-personal data and anonymized data from both government and private entities are safely accessible by research and innovation ecosystems.[24]. India and the U.S. can learn from each other’s experience to leverage anonymized data from public sources for R&D and training.

Action Item 5

Identify potential mechanisms to address some of the challenges identified in the report

»  Faculty Development: Our survey indicates that there is a dearth of faculty with expertise in AI / DS in higher education institutions outside of the IISc, IITs and a few others. Faculty development programs can mitigate this problem to some extent. Many of the NSF funded AI Institutes and Big Data Hubs offer these types of programs.

»  Access to specialized content: The dearth of faculty with core AI and Data Science expertise limits the ability of institutions to offer specialized courses. Interdisciplinary training was also seen as a major challenge for many small institutions. An immediate solution to address this gap is for Indian and U.S. institutions to co-develop content that could be hosted on platforms such as NPTEL/SWAYAM.

»  The Research pipeline: Joint Ph.D. programs and opportunities for co-mentoring of postdoctoral fellows can enhance the quality of research, increase impact, and facilitate long-term institutional collaborations. We recommend Indo-U.S. joint doctoral and post-doctoral fellowship programs, and opportunities for faculty exchange.

»  Catalyze collaborations: Funding agencies, foundations, and industry can support pilot projects that propose innovative solutions to AI and DS Workforce Development, including creation of innovative curricula, faculty training, innovative training programs for upskilling, creation of datasets, curricular materials, and case studies.

Summary

The promise and potential of Artificial Intelligence to disrupt and transform society is driving the new Industrial Revolution. The NEP (2020) is on target when it states “AI’s disruptive potential in the workplace is clear, and the education system must be poised to respond quickly.” India’s academic institutions must respond to this challenge and become global leaders in AI and DS.

The AI and Data Science ecosystems are constantly evolving with technological breakthroughs and the emergence of new application domains. The future AI Workforce will need a different set of knowledge, skills, and competencies. An Indo-U.S. strategic partnership on AI can accelerate the pace of R&D, lead to the development of global standards for education and training of the future workforce, and create a shared framework for enabling impactful and responsible innovation in AI.