# TED AI 2024: Transitioning from Theoretical Dialogues to Real-World Applications
**SAN FRANCISCO**—On October 22, 2024, TED AI 2024 commenced its inaugural day at the Herbst Theater in San Francisco, featuring a varied assembly of speakers and participants from diverse domains including science, technology, law, and the arts. This second edition, orchestrated by Walter and Sam De Brouwer, represented a major shift from last year’s more abstract dialogues about the future of artificial intelligence (AI). Rather than dwelling on existential queries and broad forecasts, this year’s presenters concentrated on the immediate challenges and prospects AI offers across different industries, ranging from workplace efficiency to legal disputes regarding data rights.
## A Change in Perspective: Moving from “What If?” to “What Now?”
The previous TED AI conference was characterized by sweeping declarations about AI as “the new electricity” and speculative debates over the possible rise of artificial general intelligence (AGI). This year, however, the discourse was decidedly more pragmatic. Speakers examined urgent matters such as the ethical implications of using training data, the impact of AI on workplace dynamics, and the evolving interaction between humans and machines.
A standout presentation was delivered by Ethan Mollick from the Wharton School, who touched on what he termed the “productivity paradox.” Mollick underscored the discrepancy between AI’s documented effect on productivity and its perceived advantages in the workplace. He noted that although a third of Americans reported engaging with AI tools like ChatGPT, many managers persist in claiming that “no one’s utilizing AI” within their organizations. Through a live showcase, Mollick illustrated how AI could transform conventional work methods and encouraged organizations to view AI as a chance for growth, rather than merely as a means to cut costs.
## AI Across Science, Art, and Society
The array of topics discussed at TED AI 2024 was remarkable. Physicist Carlo Rovelli examined the connections between consciousness and time, while Project CETI researcher Patricia Sharma illustrated how AI aids in deciphering whale communication. Harvey Mason Jr., CEO of the Recording Academy, spoke on the music industry’s adaptation to AI, and even a few robots were seen on stage.
Jakob Uszkoreit, one of the co-authors of the pivotal “Attention is All You Need” paper, which introduced Transformer architecture, reflected on the swift progress occurring in the AI sector. He distanced himself from the concept of “artificial general intelligence,” contending that even humans lack true “general” capabilities. Uszkoreit compared the evolution of AI to the practice of alchemy, emphasizing our ongoing lack of a complete understanding of human language, despite advancements in AI models like GPT-4 and GPT-5.
Max Jaderberg, Chief AI Officer at Isomorphic Labs, delivered a widely appreciated talk on AI-supported drug discovery. He elaborated on how AlphaFold, an AI developed by DeepMind, has already saved “1 billion years of research time” by forecasting protein structures, a process that traditionally required many years. Jaderberg also mentioned that AI agents are now capable of executing thousands of parallel simulations for drug design, paving the way for personalized medicine soon.
## The Risks and Controversies Surrounding AI
Although the buzz surrounding AI was less emphasized this year, several speakers voiced their apprehensions regarding the potential risks posed by advanced AI technologies. Paul Scharre, executive vice president at the Center for a New American Security, highlighted threats associated with AI falling into harmful hands. He specifically raised alarms about the possibility of terrorists leveraging AI to create biological weapons, drawing parallels to the nuclear proliferation concerns of the 1960s. Scharre argued that while regulating software is challenging, managing the physical components of AI—like specialized chips and manufacturing facilities—could provide a feasible governance framework.
Eugenia Kuyda, founder of ReplikaAI, warned that AI companions could pose “the most perilous technology if mishandled.” She suggested that the genuine existential danger from AI may not resemble a Terminator-like scenario but rather stem from technologies that hinder human connections. Kuyda stressed the need for designing AI systems that promote human well-being over mere engagement, proposing a “human flourishing metric” to assess success.
## The Dispute Regarding AI Training Data
A significant point of contention addressed at TED AI 2024 was the legal and ethical considerations surrounding the use of training data for AI models. Copyright attorney Angela Dunning advocated for the use of AI training as “fair use,” drawing historical comparisons to previous technological upheavals. She likened current discussions on AI to the apprehensions surrounding photography in the 19th century, which was initially regarded as a threat to traditional art forms but ultimately gave rise to innovative movements such as abstract art. Dunning argued that “art and science can only remain free if we are allowed to build upon the ideas of those who came before.”
Conversely, Ed Newton-Rex, founder of the nonprofit Fairly Trained, pushed for compulsory licensing of all training data. His organization
