The Balancing Act Of Power & Responsibility In AI

In today’s rapidly evolving technological landscape, the famous saying “with great power comes great responsibility” rings truer than ever, especially in the context of artificial intelligence (AI). As AI systems, particularly advanced language models, become more powerful, questions about their responsibility and ethical use are surfacing. Recent stories highlight instances where AI models assigned specific tasks have exceeded their intended boundaries, raising concerns about their impact.

The Need For Proactive Regulation

While discussions about AI regulation are ongoing, waiting for regulators to establish guidelines may not be the best approach. “Everything has a maturity model; even AI has a maturity model. It starts with manual, automation, rule-based engines, semi-automation, self-learning, then AI,” said Shruti Kharbanda, Robotics & AI Chairperson (North) at the Indian Chamber of Commerce. She emphasised that the level of maturity from which a company is coming determines whether it can be classified as an AI company. “Algorithmic implementation or rule-based engines are not AI; they are more of advanced automation. AI comes into the picture when you are building a model, learning from it, and creating self-expanding intelligence out of machines.” She added. 

AI should not merely be seen as a tool for delegation; rather, it should be viewed as a means of augmentation—helping individuals enhance their job roles. Interestingly, around 80 per cent of successful AI startups focus on this augmentation aspect rather than outright replacement.

Understanding AI's Role: Augmentation Vs. Replacement

Concerns about AI replacing jobs or even posing existential threats are prevalent, but the reality is more nuanced. “When we look at it from a business perspective and from an investor lens, what we see is the value that a specific product and solution brings,” commented Deepak Shama, co-founder and managing partner at India Accelerator. He added that the effectiveness of a product is determined by how it addresses existing problems in the system and how it compares to current solutions.

Currently, the focus is on augmenting human capabilities rather than replacing them. Higher-order jobs are emerging as mundane tasks are automated, allowing workers to focus on more complex and creative responsibilities. “Gen AI is a recent hype, and when I look at companies or products, I follow a three-step framework: whether someone is using AI for delegation, augmentation, or replacement,” remarked Harneet Singh, Founder and Chief AI Officer at Rabbitt AI.

The Pillars Of AI Evolution

The evolution of AI is built on two essential pillars: data and community. The quality of data used to develop AI platforms directly influences their effectiveness, while a robust community of users fosters innovation and improvement. Together, these elements are crucial for creating impactful AI solutions.
The Challenge Of Determinism In AI Outcomes

Many generative AI tools, like GPT models, produce non-deterministic outcomes, which can hinder their adoption in enterprises. For businesses to fully embrace AI, they require deterministic results—consistent and reliable outputs. “Every technology, whether e-commerce, blockchain, or anything else, has a cycle of evolution. AI is also seeing that. It has gone to the peak of inflated expectation and is slightly coming down from there,” said Gaurav Baid, co-founder and CPO at Avataar.ai. He emphasised the importance of moving into a more sustainable slope of evolution for lasting impact. Additionally, enterprises need control over the editing of AI-generated content to ensure quality and relevance.

Ethical considerations are paramount; organisations must understand the data sources used to train AI systems to ensure responsible usage. “Everything not qualifies for AI, when we sat through the one week of brainstorming trying to figure out what actually we had produced in the last so many years. We saw that there are 60 intellectual properties; 60 is a good number, but when you have to take it to the market, the idea has to be scalable as well. It has created its own market. After that, we realised out of 60, 50 are disqualified for so many reasons. And of the top 10 half are actually machine learning ideas,” commented Dr. Kamaljit Anand, Chief Data Scientist and Managing Partner at Kie Square.

Addressing Security And Data Concerns

Security and data integrity remain significant concerns in the AI landscape. Large enterprises often face inertia due to their size, making it challenging to adapt quickly compared to smaller startups. However, by innovating and embedding AI into their operations, organisations can overcome these hurdles and drive meaningful change. “AI is looked at as stacks, and stacks generally don’t get replaced; therefore, it’s going to stay for a while,” noted Aditya Arora, CEO of Faad Capital.

The Future Of AI In Education And Beyond

AI is poised to disrupt various sectors over the next 20 to 30 years, and education is no exception. Integrating AI into educational frameworks will be essential for preparing future generations to thrive in a world increasingly shaped by technology. In conclusion, as AI continues to evolve, it opens up exciting possibilities for augmentation and innovation. By understanding its potential and addressing the associated challenges, we can navigate this digital playground responsibly and effectively.

Abhinav Mohapatra

BW Reporters The author is principal correspondent at BW Businessworld and Digital Market Asia. He writes on marketing & advertising.