All articles/The 7 best AI TED talks in 2023
The 7 best AI TED talks in 2023

The 7 best AI TED talks in 2023

Uncover the 7 best AI TED talks of 2023! Delve into captivating discussions led by industry experts, exploring the latest advancements, ethical considerations, and future implications of artificial intelligence. Expand your knowledge and gain insights into the fascinating world of AI with these thought-provoking presentations.

Author Image

Attention all technology enthusiasts - this is your opportunity to learn about the latest AI advances and join the conversation around the ethics and realities of super-intelligent machines. Dive into this compilation of stellar TED Talks covering artificial intelligence to explore how smarter-than-human AI could improve our world. Delve into talks discussing AI innovation, future applications, as well as risks. Whether you're an AI expert eager to stay updated or simply intrigued by futuristic technology, you will discover fascinating insights here. We have curated this selection of the best TED Talks on AI to provide you with information on developments in this rapidly evolving field. Enjoy this handpicked list of presentations examining the astounding potential, along with complex considerations, in creating ultra-advanced AI.




How AI is learning what it means to be human



Walter De Brouwer argues that language evolved in humans not just to communicate but to encode the accumulated knowledge and experiences of our ancestors - effectively creating an "external library" to pass wisdom down through generations. He explains how recent AI models like Transformers have started to unlock the complex web of meaning inherent in language and give us new insight into the "celestial dance" of ambiguities, nonlinearities, and long-term dependencies that were previously hard to grasp. This provides hope that continued iterative development of AI anchored in understanding and generating human language, with sufficient human oversight and feedback, can lead to the emergence of alternative general intelligence that becomes complex, unpredictable and intrinsically human-like.




Why AI will spark exponential economic growth



We are currently experiencing the convergence of five major innovation platforms - artificial intelligence, robotics, energy storage, blockchain, and multiomic sequencing - that together are catalyzing explosive economic growth opportunities and productivity gains. According to Cathie Wood, this convergence, amplified by the rapid advances in generative AI, has the potential to accelerate global GDP growth from the current 2-3% range up to 6-9% in the next 5-10 years, while also driving a massive $200 trillion increase in the valuation of transformative innovation companies as these technologies disrupt and transform major sectors of the economy. Cathie Wood expresses strong conviction, backed by extensive research, that we are entering a highly productive and deflationary period that could grow the autonomous vehicle transportation market alone to $8-10 trillion and profoundly reshape industries across healthcare, financial services, property rights and more.




The evolution of AI — and how it will impact human creativity



In a wide-ranging discussion on the evolution of AI, Hoffman and Scott explore how recent advances, especially in natural language processing, are enabling more direct and intuitive human-AI interaction, allowing people to tap into AI's creative potential by simply using their own voice. They touch on various implications of this progress - how it shapes public discourse as AI grows more embedded in communication platforms, ethical considerations around AI's influence, and AI's capacity to augment human creativity and inspiration. The conversation also highlights how making AI more accessible and conversational opens its benefits to a broader range of people. Overall, Hoffman and Scott provide an insightful overview into how human-AI collaboration is being transformed through major leaps in natural language capabilities.




The dark side of competition in AI



Liv Boeree explores how competition, when structured poorly, can become a "Moloch trap” – a lose-lose dynamic that pushes players to take harmful actions, sacrificing long-term wellbeing for short-term wins. She shares examples like social media filters creating body image issues and news outlets prioritizing clickbait over truth. In AI specifically, She warns of a developing "rat race" mentality taking hold, with companies rushing unsafe systems to market to beat competitors. She argues this reckless competition could lead to catastrophe with extremely powerful AI. However, competition also drives progress when designed wisely. Liv Boeree urges AI leaders to "flip Moloch's playbook" by competing to implement the strongest safety measures and dedicate the most resources to developing beneficial AI. This could turn the AI race into a "race to the top” that aligns companies around the shared goal of developing safe, trustworthy AI.




Why AI is incredibly smart and shockingly stupid



Yejin Choi, a computer scientist working on AI, argues that despite impressive capabilities, today's extreme-scale AI models lack robust common sense, making silly mistakes that even children would not. She attributes this to the brute-force approach of throwing more data and compute at AI, warning it could compromise safety. Instead, she advocates focusing research on directly imparting critical human norms and common sense into AI via new algorithms and transparent, inspectable knowledge repositories. This could lead to more efficient, sustainable models that don't require inaccessible scales of data and compute. She remains optimistic about progress but believes we need more scrutiny of current methods and innovation exploring alternative approaches to instill essential qualities like fairness and ethics. Though scale has shown promise, we cannot assume it alone will lead to human-aligned AI.




The incredible creativity of deepfakes — and the worrying future of AI



Tom Graham, co-founder of Metaphysic, discusses the capabilities and challenges of artificially generated content, particularly focusing on deepfake technology. Metaphysic gained attention with a fake Tom Cruise video that went viral. Graham showcases a live demonstration, generating a deepfake of Chris Anderson's face and voice in real-time. The potential applications of this technology range from entertainment, such as creating new movies with iconic actors, to educational purposes, like having inspiring teachers speaking various languages simultaneously. However, Graham acknowledges the concerns and ethical implications, emphasising the need for individuals to own their real-world data and have control over how their avatars are created and used. Despite the weird and wonderful possibilities, he acknowledges the importance of navigating this technology's impact on trust and authenticity in a world where realistic deepfakes could challenge our perceptions.




How AI could empower any business



AI systems currently require extensive data and engineering resources, limiting their use to large tech firms. However, Andrew Ng paints a vision for democratizing AI - enabling businesses of any size to leverage its capabilities. Rather than needing huge datasets, he proposes that even a small pizza shop could provide a few key data points and receive AI-generated insights to boost decisions and profits, like predicting best-selling pizza flavors. This self-service model would remove barriers to AI adoption faced by most companies and entrepreneurs. By empowering businesses to easily tap into AI's value, we could catalyze growth across industries and communities, unlocking a wave of grassroots productivity gains that lead to broad economic prosperity rather than concentration among mega tech corporations. Ng argues that putting intuitive AI tools into the hands of local stores and startups represents a crucial milestone on the path toward an AI-powered society.



More Techbible