Rights group welcomes a policy opening
The Electronic Frontier Foundation has welcomed a California executive order on artificial intelligence as a chance to address public concern, while urging state officials to concentrate on harms that are already measurable rather than distant scenarios about super-intelligent systems.
In a statement dated September 19, the digital-rights organization pointed to biased automated decisions in employment and public benefits, AI-enabled surveillance and personalized pricing as immediate policy concerns. It called on Governor Gavin Newsom’s administration to develop rules with the participation of communities most exposed to those systems.
EFF described the order as a useful start and supported its attention to reporting requirements under California’s SB 53 for incidents involving a loss of control. The group also backed independent investigation of such events, but said California should consider how smaller developers can obtain access to third-party investigations rather than designing a process that only large laboratories can afford.
Precision around safeguards
The organization cautioned that proposed technical responses must reflect the state of the research. It said the effectiveness of so-called kill switches for advanced AI remains an active research question. Cybersecurity requirements aimed at AI laboratories should therefore be precise, practical and tied to defined risks.
EFF also raised a civil-liberties concern about any government-controlled mechanism capable of disabling a system. Such authority, it argued, could be turned against protected speech or used as political retaliation. That warning does not reject safeguards; it draws a distinction between controls operated by developers under clear safety rules and broad shutdown powers held by the state.
The statement frames California’s task as a balance between accountability, technical feasibility and constitutional protections. Incident reporting may help regulators and researchers understand failures, while independent reviews can test company accounts. At the same time, poorly specified mandates could create compliance burdens, concentrate the market or establish powers that extend beyond safety.
California has an outsized role in technology regulation because many major AI developers operate there and because its rules often influence practices elsewhere. EFF’s intervention presses the administration to use that leverage on present-day effects: whether an automated system discriminates, enables intrusive monitoring, manipulates prices or denies someone a consequential service without meaningful recourse.
The next substantive step lies with the state’s Government Operations Agency, which is preparing recommendations for the governor. EFF urged that process to include people at risk of harm and to evaluate safeguards against concrete evidence. Its position is supportive but conditional: the executive order can begin a serious policy discussion, but the resulting rules must remain enforceable, technically grounded and protective of civil rights.
