Guests Jaron Lanier, futurist, and Luis von Ahn, founder of CAPTCHA and Duolingo discuss the deceptive practices in AI systems, exploitation of crowd-working jobs, the value of personal data as labor, and the intricate relationship between intelligence, trickery, and technology.
The integration of humans and machines in AI raises questions about exploitation, fair compensation, and the true extent of our subconscious involvement in the modern AI era.
The concept of data as labor has emerged, highlighting the need to establish fair practices and reevaluate the data economy in order to balance leveraging human intelligence and ensuring equitable treatment.
Deep dives
The Chess-Playing Machine
In 1783, François-Andre Filidor, the renowned chess player, was about to face off against an unexpected opponent in Paris: a chess-playing machine called the Turk. This machine, controlled by a human hidden inside, could interact and adapt its moves during the game. Although Filidor managed to defeat the Turk, the match showcased the intrigue and excitement surrounding the possibilities of artificial intelligence, even centuries ago.
Humans in the Machine
The mechanical Turk was not the only instance of humans operating inside machines. Captchas, those squiggly letters and numbers we type to prove we're human, serve a dual purpose: not only thwarting bots, but also digitizing books. People unknowingly contribute their human intelligence to train AI systems by labeling images or providing data that improves services like Google search or voice recognition. This integration of humans and machines raises questions about exploitation, fair compensation, and the true extent of our subconscious involvement in the modern AI era.
The Illusion of AI
The mechanical Turk and modern AI share an essential characteristic: trickery. The Turk's seemingly autonomous moves amazed audiences, leading to important technological advancements. In a similar vein, modern AI uses techniques like deep learning, depending on labeled examples created by humans. As AI continues to grow, the boundary between machine and human becomes more blurred, highlighting the need to acknowledge the role humans play in training AI and the ethical challenges associated with it.
The Future of Data as Labor
The concept of data as labor has emerged, recognizing the human effort required to fuel AI advancements. People provide valuable data to train AI systems, yet often receive minimal compensation. Calls for reevaluating the data economy and establishing fair practices have arisen. As technological progress continues, the challenge lies in finding a balance between leveraging human intelligence and ensuring the equitable treatment of those working alongside AI.
In the 18th century, a device called the Mechanical Turk convinced Europeans that a robot could play winning chess. But there was a trick. It’s a trick that companies like Amazon, Google, and Facebook still pull on us today. Guests include: Jaron Lanier, futurist. Luis von Ahn, founder of CAPTCHA and Duolingo.