Are We Ready For An AGI Future?

As we inch closer to 2025, we'd like to thank you for being part of our 2024 RizingTV journey. And to celebrate a new inspired year, we're excited to announce that we're unlocking the vault to share some of our premium articles for free!

Remember November 2023? That seemed to be a tumultuous time for OpenAI. Co-founder Samuel Altman was abruptly fired over Google Meet, joined Microsoft and, then, returned to OpenAI again. Then, he fired most of the OpenAI Board that had fired him. But, just before Altman was fired, something happened.

A group of staff researchers were said to have sent a letter to the OpenAI Board. The letter was said to be about warning it about a new AI algorithm that could pose a threat to humanity. There was said to be a mysterious endeavour called Project Q* (pronounced Q-Star.) Some members believed that Project Q* could be a significant breakthrough in the pursuit of AGI, which is Artificial General Intelligence. This is a system that isn't good at one specific thing, but one that could do a wide range of things better than people. Some systems are good at something, but not everything. A smartphone could be great at understanding voice commands. But, it may not know how to learn a new language without specific programming.

AGI could create machines that aren't narrowly focused, but could understand, learn and do many things. Much like a human being. It could be a system that could learn things on its own, understand people's needs better over time and adapt to situations without needing specific instructions for each scenario. The idea of AGI is said to be building machines that have the kind of intelligence people have.

Sanjeev Menon, Co-Founder & Head Of Product and Tech - E42.ai, remarks, "The realization of AGI hinges on breakthroughs in machine learning algorithms, neural network?architectures, access to diverse datasets, advancements in hardware and interdisciplinary collaboration?among experts in computer science, neuroscience, and cognitive science".

Imagine one that's capable of reasoning and solving problems. That may make you wonder: couldn't current LLMs, like Bard or ChatGPT, already do that? These LLMs may be about to learn information and patterns from data it was trained on. But, it may not have the capability to learn from user interactions in a way that could fundamentally change its way of being. And AGI could deal with tasks including physical interaction and decision-making. While, an AI could solve problems within the scope of its programming, usually by following predefined algorithms.

So, what could that mean for the real world? If you have a virtual assistant, like Siri or Alexa, it could perform specific tasks. But, they may be limited in understanding context and adapting to different needs.?

With AGI, there could be a personal assistant that could help people with their schedule and reminders & learn preferences over time.?In the healthcare world, AGI could analyze a whole lot of medical data, understand complex medical conditions, recommend personalized treatment plans and more.

What would happen if AGI were to be accomplished in 2024?

Menon remarks, "In the hypothetical advent of AGI in 2024, AI-centric enterprises should strategically navigate this?milestone by intricately refining and expanding their AI solutions. Pivotal to this strategy is a substantial?investment in research and development, amplifying the adaptability and cognitive prowess of existing?AI models. Concurrently, fostering synergies with research institutions and tech communities is essential?for gaining profound insights into the evolving AGI landscape. Prioritizing the development of AI?applications harmonizing with AGI's human-centric nature, underscored by ethical considerations and?responsible AI practices, becomes paramount. Additionally, integrating continuous learning mechanisms?and staying abreast of AGI's technical intricacies positions these companies as avant-garde pioneers, notonly within the technical domain but also in enhancing client servicing industries".

And what could it mean for India?

According to Menon, "AGI, envisaged with human-like intelligence and human-centric design, holds the potential to?revolutionize India's technological landscape. Its integration involves the convergence of neural network?architectures, reinforcement learning algorithms and natural language processing capabilities. This?paradigm shift would necessitate comprehensive frameworks incorporating explainable AI (XAI), ethical?AI governance and continuous model monitoring.?Given India's status as a labour-intensive economy with the world's largest population, AGI could?significantly impact various sectors".

"Precision farming, facilitated by AGI-driven technologies, offers a?tangible example. With the application of AGI in agriculture, tasks such as crop monitoring, predictiveyield analysis and automated machinery control can enhance productivity and sustainability. Furthermore, AGI could address critical water-related challenges by devising efficient desalination methods, managing water resources more effectively and proposing sustainable solutions for clean water accessibility. While certain manual labour-intensive jobs may face challenges, AGI's integration opens avenues for upskilling the workforce towards managing and optimizing advanced technologies, thereby elevating the overall efficiency and competitiveness of India's workforce. The deployment of AGI in sectors like healthcare might involve advanced diagnostic algorithms powered by deep learning, while in manufacturing, robotic process automation guided by AGI principles could optimize efficiency", adds Menon.

And how does OpenAI look at AGI? OpenAI's mission is said to be ensuring that AGI benefits all of humanity. According to OpenAI, AGI could elevate humanity by increasing abundance, turbocharging the global economy and helping discover new scientific knowledge. It believes that AGI could provide people with new cognitive capabilities & act as a force multiplier for human ingenuity and creativity.

Could a single company, like OpenAI, achieve AGI?

"In the improbable scenario of a single entity attaining AGI dominance, the competitive landscape would undergo a paradigm shift, challenging the principles of collaborative AI development inherent in open-source ecosystems. Hypothetically, contemplating such an occurrence necessitates a nuanced approach. Implementing regulatory frameworks informed by federated learning and decentralized architectures could mitigate the risks of monopolistic control. Emphasizing the importance of industry-wide collaboration, consortiums and open standards becomes imperative to ensure a diverse and inclusive distribution of AGI benefits. Striking this delicate balance requires navigating the intricate intersections of technological innovation, ethical considerations and policy frameworks to usher in an era of responsible AI dominance", states Menon.

However, there may be risks of misuse, accidents and societal disruption. Surely, making a machine that intelligent could be problematic? Despite the best intentions, AGI could lead to unintended consequences that weren't anticipated during development. Maybe, AGI could still inherit and amplify biases that are there in the data it's trained on. So, there might still be discriminatory decision-making and it could reinforce existing social inequalities. And deploying AGI could be a target for malicious actors, who may be seeking to exploit vulnerabilities.

This could be for purposes, like cyberattacks, misinformation or unauthorized access to sensitive information. That could impact the integrity of critical systems. There could, also, be worries about widespread job displacement. And automation of jobs could lead to unemployment and economic challenges.

And if AGI reaches a point, where it may surpass human intelligence, how could it be controlled? Maybe, access to AGI tech may not be distributed equally. That could exacerbate digital divides, where some people or countries could have privileged access, while others may lag behind. So, there could be global inequalities that may create new power dynamics and challenges. And the AGI might interpret its goals in ways that might result in actions causing harm.

This could happen, if the system's objectives are not perfectly aligned with human values. Kind of like Ultron in "Avengers: Age Of Ultron" perceiving "peace in our time" as the death of the Avengers.

Menon opines, "If AGI becomes prepotent, surpassing human intelligence, it could bring about potential risks and?unintended consequences. Ethical concerns arise, such as a healthcare AGI deciding not to treat elderly?people to optimize overall population health. Privacy issues could be exacerbated by AGI’s ability to?process vast amounts of data. Security risks are also a concern, as AGI systems could be targeted by?malicious actors. Furthermore, AGI could lead to significant job displacement.?Despite these risks, AGI's rewards could outweigh them. It holds the potential to solve complex?problems beyond human capabilities, advancing fields like medicine, science and technology.?Responsible development with ethical guidelines, privacy protections, security measures and policies?addressing job displacement is crucial. AGI could contribute to sustainability, providing breakthroughs?like solving drinking water problems and controlling pollution. For instance, AGI might enable cost-effective nuclear fusion, addressing desalination challenges and proposing innovative solutions to?environmental issues. Additionally, AGI could play a pivotal role in tackling climate change by generating?novel strategies for carbon capture and sequestration. It may facilitate the development of?breakthrough technologies that promote renewable energy sources, reducing reliance on fossil fuels".

And could there be adequate safety measures implemented and regulated for AGI? Is that why in March 2023, there was a letter co-signed by Elon Musk and other tech experts to demand a pause in AI development?

But, OpenAI sees a significant upside, due to which it thinks it's neither possible nor desirable to halt AGI development forever. And what could happen, if OpenAI were to develop AGI? What would that time look like?

Maybe, Altman would prioritize international cooperation and collaboration to ensure the benefits of AGI are globally shared. He may form partnerships with other research organizations, governments and institutions to address challenges and opportunities posed by AGI. Or he might work to establish robust frameworks for the ethical use of AGI. So, these could involve developing safeguards, guidelines and oversight mechanisms to prevent misuse.

Could responsible deployment be ensured? Could there be a way to democratize public input and governance to avoid the concentration of power? OpenAI is said to have a principle of continuous learning and adaptation. So, maybe, Altman might deploy less powerful versions of AGI to gather real-world experience.

Is that why ChatGPT-3.5 is public? And OpenAI might have competitive pressure, which might lead to risks of shortcuts or compromises in safety measures to achieve faster progress. That might mean an AGI system deployed that may not be adequately tested. And a company like OpenAI might, theoretically, be able to manipulate economic systems or influence a political landscape, if it achieves an AGI.

Though, it's all theoretical. Even if an AI company, like OpenAI, has the very best of intentions, could an AGI deployment lead to a point where they may have no control over it? Or is that just a pessimistic and apocalyptic mindset?

"The key lies in learning from past experiences and applying those lessons to ensure the immense benefits of AGI", quips Menon.

(Become a paying subscriber of RizingTV to get access for other premium content.)