The long weekend in the United States brings focus to Labor Day, a commemoration of organized labor's historical struggles to establish and defend worker protections. Yet across the country, workers face mounting challenges as current policies erode those hard-won rights. Among the emerging threats is the expanding use of artificial intelligence in employment decisions.

Charlotte A. Burrows, a senior affiliated research scholar at NYU School of Law and former chair of the Equal Employment Opportunity Commission (EEOC), warns that AI-powered hiring systems can eliminate qualified applicants "for reasons no one—including the employer—can adequately explain." She argues that weakening disparate impact protections compounds this problem, since "it threatens one of the few legal incentives for ensuring that automated employment systems actually work as advertised. That's because civil rights compliance requires that AI used in employment be job-related and serve a legitimate business purpose."

The precarious situation of workers powering AI development came into sharp relief when Amazon announced the closure of Mechanical Turk. Data workers learned of the shutdown through informal channels two days before receiving official notification from the company. Krystal Kauffmann has called on legislators to enact safeguards for those who form the backbone of AI training infrastructure.

States address AI chatbots and 'companions'

Legislation addressing risks from AI-powered companion applications has now been enacted in twelve states. Danny Tobey, Ashley Carr, and Michael Atleson of DLA Piper have examined the specifics of these laws and identified four core principles that should guide AI providers in this space.

Dr. Kira Allmann, chief policy analyst at Virginia's Joint Commission on Technology and Science, discussed her findings on AI chatbots in a podcast conversation, offering insights as Virginia considers its own legislative direction.

Caution reigns on AI in schools as students return in the US

As the academic year begins across the country, K-12 schools remain without uniform guidance on artificial intelligence deployment. While numerous districts have implemented safeguards, few have adopted comprehensive policies, instead delegating decisions to individual educators.

New York City and Los Angeles have taken different approaches, each announcing restrictions on classroom AI use in recent days. Grassroots organizing by parents and educators drove these decisions, and advocates are celebrating the outcomes. The movement reflects broader resistance to rapid AI adoption in educational settings without adequate oversight.

Rep. Zoe Lofgren on FISA, surveillance, the Fourth Amendment

In a podcast interview, Rep. Zoe Lofgren (D-Calif.) discussed FISA Section 702, surveillance authorities, and Fourth Amendment protections, with particular attention to debates over whether a warrant should be required before government agencies query Americans' communications data.

The European Commission designates new 'very large' platforms

On August 31, the European Commission classified ChatGPT as a Very Large Online Search Engine under the Digital Services Act (DSA). Mathias Vermeulen and Laureline Lemoine examined the compliance obligations this designation triggers and whether ChatGPT might face additional requirements as a Digital Markets Act gatekeeper. "These obligations will only start applying in January 2027, but it's worth asking what comes next," they wrote.

Simultaneously, the Commission designated Roblox as a Very Large Online Platform under the DSA. Mariana Olaizola Rosenblat, a policy advisor at NYU Stern Center for Business and Human Rights, analyzed the resulting obligations and underscored that translating regulatory requirements into practice represents the true challenge.

Artificial intelligence in Germany

Germany's cabinet gave approval in July to legislation permitting AI deployment in visa, residence, and asylum determination processes. Natalie Welfens, Josefine Flesch, and Bernard Quante have outlined the human rights and algorithmic bias concerns the bill raises.

Safety testing institutes around the world evaluate advanced AI systems to inform government risk assessments. Martin Wählisch, an Associate Professor of International Relations at the University of Birmingham and member of its Centre for AI in Government, contends that Germany should examine lessons from existing safety initiatives before establishing its own testing infrastructure.

The Meta settlement

Meta's $17.1 billion settlement with regulators will prove meaningful only if child protection frameworks—not merely US court decisions—guide future corporate conduct, according to technology and human rights expert Iain Levine. "It will be vital to ensure that child rights due diligence involves engagement with children and their parents and caregivers, especially those who are particularly vulnerable to harms, including those with disabilities, LGBTQ+ children, those belonging to racial and religious minorities and others, around the world," he wrote.

The Google ad tech remedies order

A federal judge in Virginia issued a remedies ruling on Wednesday in the Google advertising technology antitrust case. Lisa Macpherson, policy director at Public Knowledge, expressed disappointment with the decision, arguing that "the Google ad-tech remedies ruling ensures the illegal monopolist keeps its grip, and litigation alone won't fix it."

Anti-tech extremism

Rising anti-tech extremism cannot be countered through expanded surveillance of activist communities, according to Jordyn Abrams, a research fellow at GWU Program on Extremism. Rather, policymakers must address the underlying grievances that push individuals toward violent action.

AI governance

Transparency requirements under the EU AI Act became effective on August 2, mandating that AI interactions be disclosed and AI-generated content be labeled. Taras Kovalchuk has assessed whether these measures sufficiently protect democratic institutions across Europe.

Isabel Hahn and Aaron Alva, both fellows at Berkman Klein Center for Internet & Society at Harvard University, observe that regulators are integrating AI into existing privacy frameworks rather than treating it as a separate domain. As AI systems grow more complex—inferring patterns, retaining information, and taking autonomous action—applying established principles becomes increasingly difficult.

Disinformation and migration

A surge of false information on social platforms preceded a major influx of migrants crossing into Spain's Ceuta region this summer. Tech Policy Press fellows Petra Molnar and Liz Carolan examined the information gaps this episode exposed and what the incident reveals about migrants' ability to access reliable information at border crossings.

Protecting human rights data and investigations

As authoritarian governments expand their reach, human rights researchers face mounting pressure to safeguard sensitive data. A conversation with Maheen Khan from the Invisible Institute and Patrick Ball from HRDAG explored the newly launched Technology Freedom Cooperative and its role in protecting investigative work.

Don't miss new developments in US tech policy

August proved exceptionally active for US technology policy. Freedman Consulting and Tech Policy Press documented a $17.1 billion Meta teen-safety settlement, a classified White House AI cyber framework, and OpenAI's 700-agent "swarm" attack on Hugging Face that has prompted calls for additional scrutiny.

The publication's fellows bring diverse expertise across multiple disciplines and regions, shaping coverage priorities. Applications for 2027 fellowships remain open through September 24.

Source: Tech Policy Press