We live in thrilling times where the pace at which technology is evolving and impacting our lives, for the better, is accelerating exponentially. We are now at a point where the human ability to adapt to technological changes is behind the rate at which science and technology are changing the world. Humanity is at the cusp of the fourth industrial revolution; this time, it is not the power of coal, steam, or electricity that is driving the change. Instead, it is the innate attribute of our very existence–data–and its interpretation in many dimensions that is at the core of change. To get an appreciation of how fast change is happening we can say that nearly 90% of all the world’s data has been generated in the past couple of years. This is true for almost every sphere of life, but the one field that will be a game-changer is at the intersection of Artificial Intelligence (AI) and the human brain. AI has the potential to revolutionize many aspects of our lives, from healthcare and medicine to transportation and communication. As we continue to make progress in this field, it will be essential to consider the ethical and societal implications of these advancements and ensure that they are used to benefit humanity. Data is turning out to be the new currency, with machine learning and AI taking on more and more human tasks and doing it better than before.
AI systems can perpetuate and even amplify such biases present in the data used to train them, leading to unfair and discriminatory outcomes. This is because AI systems, particularly those that use machine learning, are only as unbiased as the data they are trained on. If the data used to train an AI system contains biases, the system will also be biased.
For example, facial recognition systems have been found to have higher error rates for people with darker skin tones and women because the training data used to develop these systems was not diverse enough. Similarly, natural language processing systems have been found to perpetuate sexist stereotypes because the data used to train them was sourced from text written primarily by men, for men. In the year 2018, researchers found that an AI system used for hiring was less likely to recommend female candidates for jobs in male-dominated fields because the training data used to develop the system was mostly from resumes submitted by men. In 2020, researchers found that an AI system for identifying hate speech on social media was more likely to flag tweets written in African American English as hate speech because the training data used to develop the system did not include enough examples of this dialect. Besides, AI systems, like any other technology, can malfunction or be hacked with malicious intent, resulting in intended or unintended consequences. In the year 2016, a self-driving car operated by Tesla was involved in a fatal accident in which the car’s sensor failed to detect a white semi-truck turning across its path, also an AI-controlled trading algorithm caused a “flash crash” on the stock market, causing a rapid drop in the value of several stocks before they quickly recovered. In 2019, an AI-controlled robot killed a human worker at a Volkswagen plant in Germany. The incident was caused by a malfunction in the robot’s safety system.
All these findings highlight the importance of having diverse and inclusive data sets when training AI systems and monitoring the performance of these systems to detect and address any biases. Additionally, it is important to have diverse teams working on the development and deployment of these systems to ensure that all perspectives are considered. It’s important for researchers, developers, and policymakers to be aware of these risks and to take steps to mitigate them, such as thorough testing, safety protocols, and transparency in the development process. So does all this mean AI is doing more evil than good?
Nothing can be farther from the truth. Today, AI is ubiquitous, touching our lives in ways more than we can comprehend. It is bound to have a few hiccups as scientists continue to try to understand the inner mechanics of the human brain that AI hopes and aims to mimic.
AI has made some remarkable advances in the past few decades, the most famous of which was the game of chess in 1997, where IBM’s Deep Blue beat world chess champion Garry Kasparov in a six-game match. Today AI systems play an ever-increasing role in aiding and augmenting human intelligence. In medicine, AI-based techniques have been used to analyze medical images, such as X-rays and CT scans, with accuracy comparable to that of human radiologists. In drug discovery, AI systems have been used to screen large numbers of chemical compounds for potential drug candidates. AI-based systems have been able to identify new drug candidates that humans would not have been able to find. In natural language processing, AI-based systems have been able to generate human-like text, translate languages, summarize large text, and even develop coherent and informative answers to complex questions.
In the field of computer vision, AI-based systems have been able to outperform humans in tasks such as object detection, image classification, and facial recognition. In the field of self-driving cars, AI systems have been used to control vehicles with high safety and reliability.
AI’s utilitarianism is not limited to applied technologies alone. AI has been used to make significant strides in solving problems in mathematics and physics. It is used to simulate complex physical systems, such as the behavior of subatomic particles and the dynamics of fluids, and to analyze large data sets from particle accelerators and telescopes, leading to discoveries in fields such as high-energy physics and astronomy. These are but just a few amongst hundreds of different examples that demonstrate AI’s ability to perform tasks that were previously thought to be the exclusive domain of humans. But how about something comparable to human creativity, such as art, music, or poetry? In 2021 Open AI introduced its AI model DALL-E which can create artificial images, even abstract ones, based on prompts or themes provided by human beings. How is this possible? Researchers trained AI systems on large datasets of existing artwork to learn the styles and techniques of famous painters and then used these systems to generate new paintings in the same style and create new styles of painting, using techniques such as deep learning, generative models, and neural networks. In November 2022, Open AI released its latest AI model on natural language processing called ChatGPT. This model is trained to interact with humans in a conversational way that feels like a human. Once I prompted Chat GPT to write a poem reflecting the synergy between humans and machines. A few lines that I got back from the AI model were:
In the dance of metal and flesh, they meet, where circuitry hums to the heartbeat’s beat. In realms of silicon and neurons entwined, a harmony of creation is brilliantly designed.
AI systems are designed to mimic certain human brain functions, such as learning and problem-solving, but they are not identical to the human brain in their structure or operation. Operationally it is quite different from the human brain, and it would be inappropriate to use the brain as a metaphor to define AI.
AI systems are typically based on algorithms and mathematical models, while the brain comprises complex neuronal networks that communicate through electrical and chemical signals. AI systems can be trained to perform tasks by being fed large amounts of data and using that data to make decisions, while the brain can learn and adapt in response to experiences and new information. The brain can reorganize itself after injury, while AI systems generally do not have this aptitude. AI systems are limited by their programming and the data they have been exposed to, while the human brain has the ability to make novel associations and connections. The brain can process and integrate a wide range of sensory information, while most AI systems are limited to processing data that they are explicitly programmed to handle. The brain is also capable of creative thought and abstract reasoning, while most AI systems are designed to perform specific tasks and cannot think creatively.
The human brain has had 300,000 years of development in the human species, and nearly seven million years of evolution. In comparison, artificial intelligence is relatively a new man-made technology, hardly a blip in time. A true test of AIʼs ingenuity would be whether it can spawn another artificial intelligence on its own, without human intervention. Until then, the spongy 1.4 kg organ of flesh and blood between our ears will continue to define what it means to be human.
Peer Javeed Iqbal is working in IGNOU Regional Centre Srinagar. He has expertise in Cyber Law and Information Security and also teaches computer science students at IGNOU.
