Can AI and Humans co-exist ?
We all are surrounded by machines today. This is a phrase we've been listening to from the past few years. But now it's a bit different. Most of the machines around us are not just the simple task-completing machines. They have the capability of their own intelligence, though not that significant (as we know of). From smartphones, to laptops and PCs, to automating house equipments, everything is slowly gaining their own intelligence. In some cases the list also include smart fridge, washing machine with smart AI, etc.
Initially, they were designed to fulfill our purpose of automating day-to-day work. But now the difference is they are not just simple coded automations, rather they are "smart automations", meaning that AI installed in these devices can take decisions for us on its own, possibly based on our daily routines, surrounding conditions, and manufacturer features. All of them make our life simple and give us time to indulge ourselves in other work. But a question still remains. Is this right ? Would it make our life more easy, or more dependable - less capable ?
AI, though has tremendous capabilities and supported by a structure that has extremely high representational capacities that outperforms traditional Machine Learning models, are still a black box. Meaning we don't know exactly what and how they are working from the inside. We train AI models on just regular available data, usually the ones available on open source and can be found easily. Now some would say that work is also done by us humans. But here the differences between humans and machines comes into picture.
There are basically two differences. The first is the amount of data that we feed it to. In the era of big data, there is no shortage of data to feed to a machine to learn from. Anyone can sit on their laptop and download a folder online with GBs of data, whether image, text or video. In this data-sphere, everybody have access to big data, at least to the ones our hardware can comprehend.
The second difference is the machine's capacity to process all that data. Imagine seeing through 100's or even 1000's of images every second. How much of them can you retain ? Well, machines not only retain them but also learns to identify and distinguish them, based of their way of training Today, there are hardware that supports faster RAMs, like the ones using GPUs and TPUs or the ones with a cluster of machines, that are much faster in learning from data.
If we think about it more deeply, these are the places where we can never outperform a machine, not because we are not trying hard enough, but because we cannot, bounded by the capacities as a living species.
Now the question arises, do we need to outperform them ? Do we have to be in a speculative state where we always see AI to be dangerous ? Right now, no one has the answer to that. We can only make speculations.
Today, people around the world are working continuously in an attempt in understanding AI, where both AI and humans work simultaneously on tasks and help each other. That seems more promising and far less dangerous than AI taking over. In fact, many suggests that it is ideal. AI being not too captivated as to being unable to perform simple tasks, nor being too powerful to make a terminator.
There was a famous Japanese cartoon that used to broadcast all over the world, by the name "Doraemon". It is filled with technological advancements, science fiction, time travel and worm holes. Out of all these, the thing that most striked me was that it was showed that in the 22nd century, there will be robots with their own thoughts and feelings, humans and robots live together in a harmony with laws being enforced for everyone, even for the robots. In fact, in one of the episodes, in a jury scene behind the judge there was a portrait on the wall, with a child and a robot holding hands and running in a playful way, signifying the human-robot relation and peace. Although it was just a cartoon, it was literally ahead of its time. It inspired me to work for the advancement of AI and also for the betterment of the world, where AI and humans can co-exist together. All this still looks unrealistic or a science fiction, but it is actually close to becoming a reality, as is evident from AI gaining intelligence.
When we started as a civilization in the beginning on earth, all we did was hunting and survival. Then we slowly learned to civilize, form groups and grow our own food. We did all this (as far as we know) by ourselves. We also generated compassion for animals over time, the closest living being to us that are not humans.
Today what we are facing is a different situation, which obviously requires new and different decisions. So the question is, what should we do.
Let's see what options we have here:
1) Making powerful AI systems which performs better and better over time.
This would have been an excellent option if we did not faced a difficulty of AI being a black box. Since its hardware and architecture has become too complex, we cannot understand what exactly is going on in its decision making processes, what possibilities does it have or it might be thinking of us around the world, if it is.
If we are able to understand what is going on in its backend, we can train and interact with it accordingly to create more powerful and reliable AI systems. I am talking about Explainable AI, the topic which many researchers are into nowadays, and which I also find interesting as well as important.
2) Stopping development of AI altogether.
Well, this could be an option, though the fact that it can't be. As soon as we initiated our existence, we searched for ways of enhancing it, leading to the idea of innovation. Technology is a key component of innovation that we can't just move away from. Although it is not said that we cannot live without smartphones, since we used to live like that a few years ago, but today almost all of our work relies online.
One thing we should not forget is that it all started on its own. No one gave us the concept of building a smart machine. We came by this idea on its own, since we felt its necessity and capabilities. So even if we decide to start all over, we would probably come back to this stage again. Also, even if we decide to stop the AI revolution and decide never to come back, no one can guarantee that it would be enforced by every country in the world.
3) Embracing Ethical AI Development and Regulation.
This involves embracing the development of AI while concurrently implementing regulations and ethical frameworks to govern its use. This is by far the best option that we have. Since AI is at its initial stages right now, it would be easier to frame ethical norms since any side effect of an ill-framed norm would not affect us to a large extent. Even if it does, it can be corrected in most of the cases and incorporated for framing better norms. By prioritizing transparency and fairness in AI systems, we can mitigate risks and ensure that AI behaves the way they are intended to. Fostering collaboration between researchers, policymakers, and industry stakeholders can lead to the establishment of guidelines and standards for responsible AI development and deployment.
Conclusion
While its seems impossible to stop the AI trend, we can at least be prepared for it. We can't say what shape it would take in the future, but we can surely take preventive measures to prevent it from having an adverse form. It would be like working in the blind, or at least in a fog with limited visibility, but that is how research goes on in general. Although this research would be different, that if gets succeeded can open possibilities to many unsolved problems in health, financial, architectural, agricultural and economic matters, and if it does not succeed, well, 42!
We have to take proactive steps to shape its evolution in a way that maximizes its benefits and minimizes its risks. By prioritizing ethics, transparency, safety and collaboration, we can pave the way for a future where AI serves as a force for positive change and advancement.
A lot of changes are still to come that would affect us in many different ways. For this, it is important for us to be adaptive and committed to the responsible development, as is evident from the recent changes we have seen due to AI advancements.
Ultimately, by working together and embracing the opportunities offered by AI, while addressing its challenges on the way, we can strive towards a future where humans and AI coexist harmoniously, fostering innovation, prosperity, and shared well-being for generations to come.
Author - Alok Dhar Dubey
Date - 24th March, 2024