What it means to be human in the Gen AI revolution

In this new era, where AI is available to everyone and everywhere, all we need is stories with a human touch. Hence, in this blog, I am telling you my story and perspective on Generative AI, with no technical jargon, I promise!

In 2022, I completed my masters in AI & ML and then came 2023. The Gen AI revolution disrupted everything. AI which was confined only to technical people, now became available to everyone. My initial thoughts… did I do the right thing by opting for a masters, which people started referring to as traditional AI? Was it worth the investment? We are living in times when we fear machines could replace us; they are after our jobs, but what is the reality? Are the machines really that efficient and powerful that they can replace us? I bet many of us had similar thoughts. My boss when watched the video on LinkedIn that explained you can generate application code using ChatGPT. He asked me “Is this real?” At that time, I was exploring generative AI. I took time to explore and got the answer. An answer that many of us were seeking. Can Gen AI replace us?

The short answer is no! Long answer – maybe. Let’s understand this. As of now, seeing the advancements in AI and exploring it further, knowing the limitations we can easily conclude, machines can have better luck next time, now is the time for humans. We just need to be smarter than ever before. Gartner predicts that the job market, at least till 2026 will not be affected by AI. So, safe for now. Its effect is neutral. In relation to GenAI-powered chatbots, Gartner predicts that 70% of requests initiated through them will demand human oversight to validate if the chatbots are doing the right thing. This means humans are to be kept in loop, which increases the service cost by 40% because of humans and machines working in parallel. This is primarily because of the trust deficit we have for GenAI today. Stressing the importance of keeping humans in the loop, another Gartner prediction says by 2030, decisions made by AI agents without human oversight will cause $100 billion in losses from asset damage.

It sounds as if I am pessimistic about GenAI, but believe me, it’s just the other way! I am just stressing the fact that humans are important in any given scenario at least for now and few years down the line. Who knew that we could have access to music composers just on our laptop, that too without any fee? I mean it’s not bad at all. The same is true with the digital art these models can produce today, and the beauty is you just have to describe what you want. It’s similar to the disruption Google search might have done years ago, but the key at that time was you should know what you are asking! It’s the same today. All these generated artefacts may not be great the first time, but they can be refined. It can be a good starting point. This saves effort, time and energy of humans. It is a fact that we humans can get tired doing things, we would need break to recharge ourselves, but the machines can work continuously and produce things at a much faster rate than us. So, can we simply ignore that machines are becoming more powerful? If we do so, in future there are chances they can easily replace us.

So, what do we do? If machines are not replacing us (as yet) we have the option to redefine what we do, or what should we do. We should use these GenAI-powered tools to empower ourselves, enhance our efficiency and productivity. For example, a developer can use tools to generate application coding and then refining on top of that. Prototypes can be generated. Meeting notes can be taken by AI and summarized. There are so many things we can do to free ourselves from laborious and boring tasks. The industry is now talking about hyper-automation, which is combining automation and AI together to improve business decisions. So the future of work is already changing, and this is a good time(I would say already late) to redefine what we do and what can we delegate to ‘machines’.

Good, now that we know we have to redefine what we do, shall we then use machines in as many tasks as we can? Life would be easy if we can do that. But there is a glaring problem with solutions born from the Gen AI revolution. They lack the ability to explain the decisions they make or the outcomes they produce. The intelligence they have is based on the data that has been fed to the machines. If the data quality is bad, their decisions may be of bad quality. If the data is biased, the outcomes can be biased as well. Taking about security these machines can bypass security features like captchas. Machines trained on copyrighted material could infringe on intellectual property rights. It can be dangerous too as creating fake content can have a very negative impact on society – deepfakes are already in the news. There could be ethical implications leading to unintended consequences. It has a considerable impact on the environment as well because these machines are resource-hungry, they need a significant amount of energy to run producing significant carbon emissions. The point I want to make is that the machines have issues as well and hence it should always be evaluated if we need machines for a particular use case. Not only technical but ethical evaluation too.