Debjani Ghosh on Importance of Human Choices in an AI World

2023 has emerged as one of the next big frontiers in human-artificial intelligence (AI) collaboration. However, as we explore the possibilities and consider the challenges, a few fundamental questions remain unanswered: can India’s tech services industry expect significant revenue growth from this technology soon? What are the risks associated with large-scale deployment of generative AI? Why is human control crucial for the future of AI?

AI has frequently been employed to augment analytical tasks, aiming to boost productivity or attain significant advancements. For instance, oncologists have leveraged machine learning techniques trained on extensive datasets and medical images to drastically enhance the precision of cancer detection. AI augments their capabilities, elevating the significance of experience and specialised knowledge in the process.

Insilico, an AI-driven drug discovery company, advanced to Phase 1 clinical trials within two and a half years of starting a project targeting a rare respiratory condition. It is now progressing to Phase 2 trials. Following traditional methods would have taken over six years and cost upwards of $400 million for the same endeavour.

Magnifying this to the tech services industry, as per our recent report with McKinsey, generative AI is projected to create an annual economic value of $2.6 trillion to $4.4 trillion, with roughly 75% concentrated in key areas like software engineering, customer operations, product development, R&D, sales and marketing. These functions form the core of the Indian technology services industry.

We have pinpointed four crucial areas where generative AI is revolutionising and advancing multiple facets within service offerings for the technology services industry.

Expansion in market reach: The emergence of over 100 generative AI applications has the potential to drive an additional 15% to 20% growth for leading providers by introducing new or reimagined offerings within the next five years.

Enhanced delivery efficiency: generative AI is poised to enhance delivery productivity by approximately 30% within the next two to three years. Notably, significant productivity gains are expected in specialised service lines like application development and BPM services.
Sales optimisation: Anticipated productivity gains span the entire spectrum of sales and marketing activities, from boosting lead generation to facilitating quicker formulation of sales strategies.

Increased productivity in G&A: over the next three years, an estimated 40% boost in productivity, including sales, is foreseen through task automation and augmentation across functions such as finance and accounting, legal and HR. However, the potential for impact depends on the service mix, the concentration of tasks across complexity levels, the vertical focus and the posture taken by the service providers.

Power of Collective Choices

Amidst the hype and excitement surrounding AI’s capabilities, a critical question arises: who is truly shaping the narrative of AI’s development? As AI systems advance in sophistication, the escalating concern regarding their potential risks becomes increasingly alarming.

It is tempting to perceive AI as an autonomous force shaping our lives, but the truth is more complex. We, as humans, are not mere bystanders in the progression of AI. The decisions we make—both individually and together—will profoundly shape the trajectory of AI and, in turn, define our own future.

Contemplate the evolution of facial recognition technology—a potent tool with the potential to bolster security, yet fraught with substantial risks of discrimination and privacy breaches. The choices made throughout its development, particularly concerning data collection, algorithm design and deployment, will ultimately dictate whether it emerges as a force for societal benefit or as a tool facilitating oppression.

Demanding Accountability

AI cannot be treated as an isolated entity, detached from our actions and their consequences. This demands a fundamental shift in perspective, advocating for a human-centred approach to AI. Such an approach underscores the imperative for:

Transparency: it is crucial to validate algorithms using additional mechanisms before deployment. Testing algorithms for unintended consequences stemming from subjective inferences or biased data is essential to ensure their reliability and fairness.

Human oversight: the data employed to train AI models might inadvertently contain implicit biases tied to racial, gender, origin or political identities. Scrutinising data input into AI systems for these biases is essential, as they can distort algorithms and subsequent outcomes. Intervening with human judgment is vital to ensuring that ethical considerations guide AI usage.

Building safe AI instead of making AI safe: rather than aiming to secure AI post-creation, a pivotal shift would be to craft “safe AI” right from its inception. This proactive stance emphasises embedding safety measures throughout the entire lifecycle of AI development, spanning from its conceptualisation to deployment. Creating safe AI demands a collective effort by engaging researchers, developers, policymakers and the public, all working together to ensure that AI serves the best interests of humanity.

Regulation as a shared responsibility: Regulation represents a shared responsibility between industry and government, necessitating a three-tier collaboration for effective oversight. Initially, industry self-regulation is key, requiring companies to proactively demonstrate responsible AI development, emphasising transparency and accountability.

At the national level, regulations should foster AI’s positive impact while safeguarding against risks, promoting innovation within borders. International cooperation becomes essential due to AI’s global reach. This collaborative approach ensures that AI evolves globally while adhering to fairness and ethical principles.

Looking into the Future

As the technology landscape advances, the question arises: When will major technology services firms in India start seeing tangible monetisation gains from deploying generative AI? Technology service providers must strategically approach their future success by looking through both short- and long-term lenses.

In 18 to 24 months, a microscopic view can focus on tactical steps like resourcing, tooling and use case validation. Over the next three to five years, a telescopic view can dictate broader organisational shifts to mobilise the ideal operating model and value proposition.

In conclusion, AI stands as a driver for economic advancement with the potential to revolutionise industries and elevate the quality of life. To fully unleash its capabilities, embracing pro-innovation regulations that prioritise ethical standards, transparency and collaboration between industry and government, both nationally and globally, is imperative.  By recognising our influential role in shaping the AI narrative, we pave the way for a future where human-AI collaboration thrives. AI can be a catalyst for progress, tackling intricate global challenges and enriching lives worldwide.

Active participation in steering AI’s development ensures it remains a force for good, dedicated to serving humanity’s best interests. Let us not merely observe but actively engage in architecting a future where AI enriches our lives and fortifies our collective human journey.

The author is the president of Nasscom, the apex body for India’s technology industry