Dear Member of Congress,
On behalf of the AFL-CIO, representing 15 million working people across 63 affiliate unions, we strongly urge you to oppose the “Strengthening Artificial intelligence Normalization and Diffusion By Oversight and eXperimentation Act” or the SANDBOX Act (S. 2750). This bill, recently introduced by Senator Ted Cruz, allows companies that develop and deploy AI tools to apply for waivers that exempt them from federal regulations for AI testing. With no comprehensive federal regulations in place for AI, the SANDBOX Act would exempt companies from any future AI regulations, as well as other regulations swept up in the overly broad definition of AI, for up to a decade. Lawmakers must oppose any policy that shields big tech from government oversight, whether it is through a waiver process or any sort of federal preemption measure.
The SANDBOX Act represents a dangerous overreach by the technology industry. At a time when the Big Tech lobby is attempting to ban states from regulating AI, it is now imagining new ways to ensure there are no guardrails in place at the federal level either.
Under the guise of accelerating innovation, the Cruz SANDBOX Act all but ensures that technology companies, and the companies that use these AI systems, will be able to write their own rules, leaving workers and the general public with no protections against the well known harms from unregulated AI. Additionally, because of this bill’s broad definition of artificial intelligence, it may also exempt companies with waivers from existing regulations regarding other technologies.
Data-driven technologies, like AI and machine learning will dramatically reshape work, and without proper regulation, could create significant harm for working people. Already, we are seeing new technologies being introduced to workplaces, undermining workers’ rights. Across a myriad of industries, employers are using untested, unregulated automated decision and algorithmic management systems for key functions, including hiring, scheduling, and setting wages. These technology systems are often linked to negative worker outcomes, including increased psychological stress, injury risk, scheduling and income instability, job elimination and deskilling, burnout, and turnover. In some cases, the implementation of data-driven systems impacts compensation structures by "deskilling" work, depressing wages, eroding job security, or undermining royalty structures by threatening essential copyright and intellectual property protections.
Congress has stalled on taking meaningful action to establish safeguards in the public interest. Now, many members of Congress are ascribing to a dangerous narrative that for the US to lead in AI innovation, they must remove all regulations so that Big Tech can race untested products to market. We are concerned that embracing this reckless approach to deployment at the cost of safety and sustainable innovation is leading to dangerous experimentation yielding severe and avoidable outcomes across the private and public sector. Without safeguards, cars are put on highways equipped with untested self-driving systems that don’t recognize or respond to railroad crossings, leading to collisions and even tragic but avoidable deaths. Workers’ data are used to subject them to relentless productivity quotas facing injury or termination at the behest of an algorithmic management system. In healthcare, the unregulated use of AI-enabled technologies could pose severe risks to patients. And, this is the tip of the iceberg, if our federal AI legal framework upholds the interests of Big Tech to the detriment of working people and the general public.
The SANDBOX Act is another attempt to undermine transparency and accountability for AI development. Under the bill, companies developing AI systems would self-assess whether their product presents a “health and safety” risk, a risk of “economic damage,” or a risk of “unfair or deceptive trade practices,” but definitions for these risk assessments are narrow. Health and safety risks account for physical harm to consumers, but not mental or emotional harm or any health and safety risks posed to workers. Economic damage only protects property and assets of consumers, not wages, benefits or job security for workers. Further, requirements for assessment risks do not include discrimination and disparate impact. There is no recognition of algorithmic bias, disparate treatment in hiring, firing or other adverse actions.
This bill comes after a failed attempt by Senator Cruz and other Republicans to halt state AI regulations. The House passed the proposed moratorium earlier this year as part of the “One Big, Beautiful” bill (HR 1). Despite intense bipartisan pushback, and Congress resolutely rejecting the moratorium, President Trump’s recently released AI Action Plan attacks state-level AI regulation by threatening federal funding in a clear attempt to advance the moratorium agenda of Big Tech
Efforts to shield tech companies from existing or future federal regulations, or enact federal preemption provisions are out of step with the demands of the general public. A recent Gallup poll found that “80% of U.S. adults believe the government should maintain rules for AI safety and data security, even if it means developing AI capabilities more slowly.”
Lawmakers have an obligation to listen to Americans and enact an AI framework that promotes responsible innovation through strong worker protections, enforcement measures and accountability standards. The SANDBOX Act shields Big Tech from government oversight and blocks lawmakers from responding to real-world harms that come from unleashing unregulated AI systems into our communities. Lawmakers in the states are doing their part to enact robust safeguards, developing bipartisan legislation that protects workers from surveillance and algorithmic management, minimizes deepfakes and disinformation that undermine safety and public trust, and erects robust guardrails that prevent dangerous experimentation with AI-enabled automation systems.
We urge you to act in the best interest of workers and the general public in pursuit of responsible innovation, and oppose Cruz’s SANDBOX Act and any legislation that shields Big Tech from government oversight at the federal, state, or local level.
Sincerely,
Jody Calemine
Director, Government Affairs