AI Now Institute argues AI deployment can deepen inequality and concentrate power

AI Now Institute argues AI deployment can deepen inequality and concentrate power

AI Now Institute published this chapter on June 3, 2025. It argues that AI’s promised benefits are overstated and that its deployment in public services and workplaces can worsen people’s lives, weaken expertise, and concentrate power. The chapter’s central point is that AI is not just a tool whose value depends on technical performance: institutions choose where to use it, whose knowledge to rely on, and who bears the consequences when it fails. The Institute links AI adoption to budget cuts and weakened public services, describing examples involving schools, healthcare, government benefits, housing, and immigration enforcement. It argues that workers often face surveillance and worsening conditions rather than sharing in productivity gains, while people subject to automated decisions may have little transparency or way to challenge errors. The chapter also disputes claims that prospective breakthroughs, such as AI helping cure cancer, justify unchecked industry growth; it notes that medical advances still require human work and clinical testing. Its conclusion is that flawed systems can still serve the interests of institutions and companies deploying them, while harms fall on people with less power. The chapter does not establish how often these effects occur across all AI deployments, and it says MyCity’s fate remains undetermined.

Publication: 0462b571-965d-7dfc-6ce6-f180ce9797e6

attribution — pass: The reading attributes the chapter and its arguments to the AI Now Institute, the publisher of this single report chapter. Its claims are framed as what the Institute argues or says, not as established fact.

In this section, we describe how the tech industry has sought to reshape society

summary — uncertain: The reading softens the source's conviction. The source says AI is consistently deployed in ways that make lives worse, and the chapter is titled as AI consistently failing the public. The title and summary instead say deployment 'can deepen' or 'can worsen' outcomes. The added caveat about frequency further dilutes that force. Smallest repair: attribute the 'consistently' claim directly, for example 'argues AI is consistently deployed in ways that make people's lives worse'. Minor point: the closing mention of MyCity does not say what it is; the claims identify it as New York City's portal.

It is consistently deployed in ways that make everyday people’s lives, material conditions, and access to opportunities worse.

date — pass: The stated date of June 3, 2025 matches the UTC calendar day of originalPublished (2025-06-03T06:00Z). Dates cited in the text, such as a May 17, 2025 access date, precede publication. No later events or hindsight framing appear. This is a publisher page, not an archive copy, so no archive disclosure is needed. The retrieval-date limitation is accurate.

support — pass: Each claim is source-stated and its excerpt matches the evidence text. The statements that medical advances require clinical testing and that flawed systems still serve deploying institutions are supported by the source. The question is open-ended and follows from the source's concerns about public services.

AI systems, even when faulty, act in the interest of enforcers who need to meet targets for deportation and arrest

media — pass: No media was selected and no verified media candidates were supplied, so there is no identity or relevance issue.