AI: The next revolution has begun

Introduction

Just over a year and a half ago, we were amazed to see ChatGPT in action for the first time. I’m sure we all tried out different prompts, and it left us amazed when ChatGPT churned out content, phrase after phrase, without missing a beat. It seemed to have an answer or an opinion on anything and everything.

As Director of Technology at 99x, I have spent 20 years adapting to numerous shifting technology trends—from classic desktop-based applications to web applications, cloud-based technologies, and mobile applications. There have been many technology shifts, some longer lasting than others. Many emerging technologies were tagged as ‘game-changers’ but failed to impress and fizzled out over time.

However, I soon realised that ChatGPT was different. As a company, we realised that the underlying generative AI technology behind ChatGPT had the potential to impact our lives more significantly than we could imagine. It had the potential to weave into our daily activities seamlessly, just as we experience interactions with navigation and location-based services today.

We also understood that the development of a generative AI-centred technology shift was as fundamental as the creation of the microprocessor or the mobile phone. It will change the way people work, learn, travel, get healthcare, and communicate with each other. Entire industries will reorient around it, and businesses will distinguish themselves by how well they use it. Bill Gates recently remarked that he considers OpenAI’s GPT model to be the most significant breakthrough in technology since he first saw the graphical user interface in 1980.

Beyond clever algorithms

Although AI has been a buzzword for a long time, I’ve been reluctant to use it too often as people tend to misuse the term. At least in the past, AI adoption required a massive investment and effort to implement, but with inadequate returns. When someone asks us about the AI strategy for 99x, there’s no simple answer. That’s because what most people call AI is often just clever algorithms or complex statistics. For example, take a system which can predict the sales forecast for a product based on area or time of day. The results could be accurate and impressive, tempting someone to call it an AI solution. While developing a system like this could be complex and may deal with large volumes of data, it is not true AI.

Co-pilot based applications

With the advent of large language model-centred generative models, we can create human-like co-pilots that can assist users in ways that were not possible before. This can significantly improve the productivity of complex, software-driven workflows while reducing errors. This increased level of productivity was hard to achieve in the previous era of AI and software technologies.

During the past decade, software applications have successfully digitised complex processes in domains such as manufacturing, finance, and healthcare. However, the application of business logic in these applications had reached a plateau of improvements that could be made to a product. Recent research at 99x Labs revealed a new paradigm that can bring massive productivity gains. This was done by integrating natural language-based workflows to run alongside existing applications. As a result, we have seen significant improvements. This approach enhances productivity and simplifies the user experience, allowing us to recompose older, complex applications more effectively.

In this new AI co-pilot-driven approach, applications are transformed to become more human-like. Large language models (LLMs) enable dynamic code generation and the execution of new tasks that were not originally hardcoded into the applications. This allows applications to think and rationalise actions, greatly enhancing their functionality and adaptability. The possibility of transitioning into these advanced architectures is becoming increasingly feasible. This allows us to move beyond overhyped, imaginary concepts into practical, real-world applications which ultimately add business value and benefit people.

Improving explainability and transparency

In traditional AI models, the primary focus is often on achieving a high degree of prediction accuracy. However, these models are typically black boxes, making it difficult to understand how they arrived at their conclusions. Improving the explainability of factors contributing to a model’s accuracy is far more important than the result. For example, if your model can accurately predict liver disease, is that enough for the doctor or patient? Wouldn’t you want to know exactly why someone was placed in a higher-risk category? What were the contributing factors?

In our new approaches, we are investing heavily in research and expressively generating the core factors contributing to the patient’s current condition. This approach enables doctors to plan treatments more effectively and manage a patient’s health to prevent further progression of the disease.

The road ahead

 We are fortunate to witness and experience this revolutionary wave of technology that will drastically change the world as we know it. These advancements will solve many challenging problems that were difficult to address due to the limitations of past technologies. It is crucial to embrace this change rather than be intimidated by the threats and fears associated with such drastic technological shifts. I’m reminded of the quote, “While AI won’t take your job, a person more proficient in adopting AI could.” The same applies to business as well. Businesses that accept the new AI-driven solutions will thrive. In conclusion, we must respond by seeing this as an opportunity-based change cycle rather than a fear-based one.

(The writer is 99x Director of Technology.)