The enormous resource demands of AI data centers discussed in part one of this conversation raise an important question: what is all the computing power ultimately going to be used for?
Hakeem Anwar examines the claims of an “AI arms race,” the potential for increasingly pervasive surveillance, and the dangers of dependence on centralized technology. He also presents another path—private devices, locally hosted AI, and practical steps individuals and communities can take to retain greater control over technology.
An Increasingly Powerful Surveillance System
Smartphones, computers, cameras, smart speakers, televisions, and other connected devices already collect enormous quantities of information about everyday life. Anwar warns that AI could dramatically increase the ability to combine, analyze, and act upon this information.
He describes a plausible future in which surveillance becomes increasingly centralized and automated, with people’s activities across numerous devices feeding into unified systems.
Are We in an AI Arms Race?
Anwar describes the rapid expansion of AI computing alongside claims of an AI arms race, particularly between the United States and China.
He argues that consumer and business demand alone cannot convincingly explain the scale of the buildout. Citing an MIT report, Anwar says that around 80% of China’s data centers are not being used, despite the country’s extensive adoption of AI, robotics, self-driving vehicles, and other AI-powered technologies. If China can sustain such widespread AI use without approaching the limits of its existing capacity, he questions why vastly more infrastructure is supposedly required in the US.
This leaves an important question: what is all the additional computing capacity intended for? Anwar raises military and national-security applications as one possible explanation, bringing with them their own set of serious concerns.
Another Way toincreasingly Use AI
Rejecting centralized AI does not necessarily mean abandoning AI altogether. Anwar advocates for truly sovereign AI: open-source models that can run locally on a person’s own hardware rather than sending their data to remote servers over the internet.
Some of these models can operate entirely offline, allowing information to remain on the user’s device. As we increase the capablility of open-source models they can become an alternative to dependence on centralized services.
Removing Surveillance From Your Devices
AI expansion is only the most recent privacy issue. Phones, laptops, televisions, and other connected technology can already collect and transmit information about their users. People can begin learning how to important to remove unnecessary AI and surveillance layers from phones and computers, while switching over to privacy-focused alternatives.
Local Communities Can Push Back
Individual technological independence is just one way to prevent this overreach. Decisions over new data-center projects are often made locally, giving organized communities an opportunity to intervene.
Anwar points to San Marcos, Texas, where organized opposition challenged a proposed $2 billion data-center project. Residents held regular gatherings and brought around 100 people to a planning and zoning commission meeting, where they continually made their voices heard.
Drawing a Line With AI
The choice is not simply between embracing AI and abandoning technology altogether. It is about finding ways to use make it ethical, private, and a healthy part of our lives. As centralized AI becomes more deeply integrated into everyday life, individuals will increasingly have to decide where to draw the line.
Learning to use private devices, locally hosted AI, and decentralized alternatives can help preserve greater technological independence before reliance on centralized systems becomes too difficult to escape.