For those who are ardent believers in technology and its power to make the world a better place, artificial intelligence means everything we’ve been told we can expect from what has been described as a “brilliant future.” There is a possibility. For the novice among us, it can mean the abstraction of an idea, something powerful and pervasive without form that gives machines the power to do everything. It helps you find a faster route to work, act as a therapist with human empathy, and anticipate.
Opponents, on the other hand, support vigilance, arguing that unregulated deployment of AI technology will cause irreparable harm. Popular culture has consistently portrayed this doom story. Joan is Awful, the new Black Mirror season opening episode deals with problematic ethical considerations for AI. Here, a streaming her platform collects details about Joanne (Annie Murphy) and, with the help of quantum computers, produces a daily show that mirrors her life. The platform does not seek explicit consent from either Joan or actress Salma Hayek, who plays the AI-generated version of Joan in the show.
Aside from the imagined futures and tech dystopias, the concerns are immediate. The UN Security Council is scheduled to meet this week to discuss the threat AI poses to international peace and security. Secretary-General António Guterres recognizes the importance of AI in achieving the Sustainable Development Goals, but intervenes to prevent AI from belittling human agency and evolving into a “monster that no one can control”. asked for
According to experts involved in the implementation of these technologies, in the face of breakthrough technology and unprecedented innovation, the way forward is to build guardrails and prepare with an open mind.
Responsible recruitment is key
Calls for state regulation of AI technology are gaining momentum in India, but industry and researchers are supporting a less conflicted approach through policies that incorporate responsible AI thinking into sector-wide adoption. seems to be While there is consensus on the urgency of the issue, including the threat of privacy violations, stakeholders also agree on the need to nurture these technologies as they evolve and stabilize, choosing preparedness over outright resistance. It also emphasizes gender.
Professor R Venkatesh Babu, Department of Computational and Data Sciences, Indian University of Science, emphasizes the incredible pace of adoption of AI technologies across disciplines. Research, both institutional and corporate-funded, examines the range of these technologies, and industry spends its time applying its findings to products and services. “For AI research, there is no gap between academia and industry,” he says.
Monitor but do not over-regulate
While generative AI technologies raise concerns, many of which relate to issues of inherent bias, some scholars argue that in order to guide these technologies into their most commonly applied forms without over-regulating them, , urges us to monitor these technologies closely. “Models like ChatGPT are trained on huge amounts of data. They can also be thoughtful, meaning they will learn and improve on their understanding of the data, but the problem is that these systems also contain biases from the datasets they use. These biases can be further amplified,” says Professor Bab.
AI systems can inherit these biases from data undersampling and discriminatory datasets. Distorted inferences, when used to help humans make decisions with their own implicit biases, lead to misrepresentations based on gender, race, and social indicators, resulting in less accurate outputs. There is likely to be.
This contradiction is consistent in AI applications in art forms that reinforce cultural and ethnic stereotypes, immature algorithms used in job recruitment, facial recognition systems that fail to classify dark faces, and clinical diagnostics. are appearing. Based on datasets lacking patient diversity.
break down prejudice
This bias manifests itself in important areas such as healthcare. Algorithms dealing with region-specific data can lead to inconsistent and inaccurate results when analyzing disease prevalence outside the population of a given region. Dr. Arjun Kalyanpur, Chief Radiologist and Founding CEO of Teleradiology Solutions, believes curation and tailoring of data to meet specific requirements is essential to the sustainable adoption of AI technology.
Teleradiology Solutions develops algorithms that power NeuroAssist, a product that helps instantly distinguish between hemorrhagic and non-hemorrhagic strokes, allowing clinicians to make course of action decisions even when a radiologist is not on site. lead to Another of his products the company has developed, his MammoAssist, studies mammogram data to identify subtle patterns in early-stage breast cancer.
autopilot analogy
Dr. Kalyanpur stresses the overwhelming shortage of radiologists in India, arguing that there are only about 20,000 radiologists and doctors need all the help they can get. AI applications are coming at the right time, he said, and some concerns about it may be overstated. “I do not believe that AI systems should replace humans. You’re in control, but in autopilot mode you can do some of the tasks that are mechanical and repetitive in nature,” he says.
Black Mirror showrunner Charlie Brooker recently said that the series (previous series were entirely set in the not-too-distant future) is a pernicious accusation of technology that has been interpreted so far. said to be completely different. It may be the fault of those who are not coping well. This has always inspired literature and art, but it is the result of human involvement in change.
Thus the cyborg artist
AI is democratizing art, launching a platform where graffiti artists and amateurs can create images that make up an entirely new genre. AI systems like Dall-E, which generate realistic images from prompts, despite raising concerns about ownership (whether the art belongs to the user or the platform) and lack of original ideas , such images are also generated from those already created. But for professionals, AI will become an experimental tool for pursuing higher representations.
“I consider myself a cyborg artist,” says KK Raghava. In 2018, the multidisciplinary artist, along with his brother Karthik Kalyanaraman, launched the art curation and research collective 64/1, which aims to advance and build public understanding of artist-AI collaboration. ” was established.
“The question to ask is not whether AI is good or bad. The question is whether we have such an imagination for this country. We need artists and innovators to create this alternative imagination, and the West will always have something we don’t have: science fiction. I did,” says Raghava.
deal with the inevitable
Bangalore-based emotional AI startup Entropik has introduced face coding, eye tracking and voice AI technologies to help brands understand consumer preferences. Ranjan Kumar, CEO of the company, said Entropik’s multimodal technology interprets emotions such as happiness, sadness, anger and surprise, providing insight into consumers’ emotional reactions. For example, Speech AI evaluates features such as tone, pause, and pitch to identify emotions in speech and rate them as positive, negative, or neutral.
Kumar says the technology will allow companies to go beyond traditional research-based methods to gain faster, unbiased insights. “AI enables businesses to make data-driven decisions, tailor products and services to customer preferences, and create more engaging experiences,” he says. .
As ChatGPT urges educational institutions to phase out online exams, startups identify AI-powered solutions to mitigate climate change. As a series of copyright infringement lawsuits hit AI platforms, courts are also using his AI-enabled transcription service. “It is important to emphasize that AI is not intended to completely replace human jobs, but rather to enhance human capabilities and create new opportunities,” Kumar said. . It’s important that he uses the word “fair” to qualify his company’s performance. It essentially adds context to discussions about impending and inevitable change and the various ways to handle it.
Raghava sounds ready when he says, “We can’t stop the storm. All we can do is adapt.”
Here’s a quick primer
* Artificial intelligence (AI) is a type of simulated intelligence that mimics human capabilities and trains machines and computers to perform specific tasks.
*Part of the larger AI spectrum, Machine Learning (ML) involves the use of models that study large amounts of data and learn and evolve over time while improving the accuracy of the output. ML systems are widely adopted as processing large amounts of complex data becomes increasingly important and beyond human capabilities.
*Some AI applications run on deep learning systems, hierarchical neural networks inspired by the human brain.
* Tech giants and corporations are investing heavily in AI technology. Extensive research also complements the adoption of these technologies and continues to diversify through their subsets and specific application areas. For example, using dialogue datasets that can train AI agents (chatbots) to be more ‘understanding’ when providing mental health counseling, or tracking eye movements to better classify a subject’s level of engagement. Etc.
* AI systems have found innovative applications across a variety of domains, from coding to art to precision agriculture. Generative AI, a collective term for systems that create new content from existing data, got its breakthrough moment with ChatGPT, a model that responds to user queries by culling data and generating relevant text.
* Threats posed by AI-derived products like deepfakes (faking or manipulating visual and auditory content) are being addressed through efforts to identify and perfect detection techniques. For AI to be an enabler, experts say, these processes of constant checks for evolving threats must be non-negotiable.
