eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
AI scholar and professor emeritus at New York University Gary Marcus has called for regulation to protect democracy as improving generative artificial intelligence models increasingly become potent tools of lies and political propaganda with deepfakes and misinformation. In his latest book, Taming Silicon Valley, Marcus incisively uncovers the surface layers exposing the present and future dangers of generative AI and lists its real-world effects.
The Danger Marcus Sees
Generative AI is being used to spread lies and misinformation, imitate copyrighted books and works of art, and proliferate cybercrimes and scams in the virtual space and on social media. There’s also increasing concern about consumer data privacy in the hands of unregulated tech companies that are free to sell private data.
In the eyes of many experts, including Marcus, the U.S. is a digital Wild West. The danger that technology poses is not so much about a doomsday scenario in which AI annihilates humanity as what large language models like ChatGPT have already done to society and what further harm it will cause to people and institutions when left unregulated.
“(There) is a lot of talk of… the probability of machines annihilating all people,” Marcus writes in Taming Silicon Valley. “I doubt that we will see AI cause literal extinction.”
The Solution Marcus Proposes
In an interview with The Register about what is most troubling about ChatGPT and similar models, Marcus cited “the threat to democracy from automatically generated misinformation and deepfakes.” He suggests that a regulatory body—an AI agency that delves deeper into guidelines and rules—should be formed to regulate and promote the responsible use of AI and digital technologies in general.
He decries big tech companies “becoming increasingly powerful, with almost no constraint,” and argues that Silicon Valley must be tamed. For that, public pressure is needed. “Without more public pressure,” he said, “very little will be done in the U.S. to protect citizens from increasingly invasive and problematic AI.”
AI-generated deepfake voices, videos, photos, and books being used to scam people and big tech companies acting without accountability for their actions are just some of the dangers posed by artificial intelligence, and they are already harming society. AI has already been used to create and disseminate disinformation that has influenced elections in the U.S. and the Philippines. Authoritarians thrive when the truth is compromised, and alternative versions of reality replace facts. Without serious legislative work to establish regulations, generative AI endangers America’s much-cherished democracy.