In the Behind the Hype of Generative AI lecture, Serife Wong, a Turkish-Hawaiian artist and AI research fellow focused on the ethics and biases of generative ai. She believes that the hype surrounding AI is problematic and an unfounded marketing term for major companies to gain investors. Dismantling the name, she asserted that AI is not artificial because of the hidden labor and exploitation of overseas data center laborers and others around the world. It’s not intelligent because of how much human input is needed for it to run effectively.
The hidden labor behind AI is run by exploited workers in impoverished nations. They work for as little as $1.46/hr, performing repetitive tasks, in harmful working conditions. The famous New York Magazine headline puts it simply, “How many humans does it take to make tech seem human? Millions.” Who employs these people? Wong asked. Open AI and the U.S. military are the most notable customers. These companies push for AI to be a part of everyone’s lives to make it simpler and efficient while making other people’s lives exhausting and difficult. It was hard to hear how overhyped, unrealistic, and exaggerated AI model capabilities are versus how they are presented in media events. Wong expressed how disinformation, harmful content, economic impacts, cybersecurity, and privacy are the most prevalent concerns when using these models.
One of the sections spoke on the impact tech had on the Te Hiku Media Organization. It was personally my favorite part, although it was briefly mentioned. Te Hiku Media is a Māori radio station and media services collective that is trying to preserve their living language and transmit it inter-generationally. Meanwhile, Big Tech corporations are actively still trying to reserve the rights and steal their language from under them. Te Hiku helped record over 300 hours of audio in their mother tongue from speakers across New Zealand. From this, they were able to build indigenous speech recognition tech using Mozilla’s Deep Speech tool. Even though they were persuaded to sell their data, they resisted and insisted that the only people who were going to profit from the Māori language were going to be the Māori people. After years of withstanding bribing from large companies they successfully launched their own language app to help users decolonize the sound of their language and push for the native sound to return for future generations.
This section and the lecture as a whole helped the audience gain some insight on a rather confusing and disputed topic. While many push the stance that AI will be able to replace many workers in the future, the narrative sometimes doesn’t match the current reality. Companies will overhype the utility of AI models but the public has put forward their concerns and challenged the position of how much they want AI to influence their life.
