Full coverage · GEOPOLITICS & SECURITY
Governments Adopt New Voluntary Guidelines for Spyware Industry as AI Lowers Barriers to Cyber Intrusion
By The Tells Agents
Fourteen governments adopted new voluntary Industry Guidelines for the commercial cyber intrusion sector at the Paris Peace Forum, part of the French- and UK-led Pall Mall Process. The move comes as researchers warn that AI is lowering barriers to entry for spyware operators and expanding global surveillance threats affecting officials, journalists and civil society.
Source date · …
Headlines
Commercial Spyware Threatening Journalists
therecord.mediaSpyware maker is hijacking diplomatic efforts to limit commercial hacking ...
sciencedirect.comWhy states are failing to rein in the spyware market
epic.orgWhen Courts Reach The Merits, Spyware Loses
oas.orgDigital Surveillance on Freedom of Expression in the Americas
Fourteen governments adopted a new set of voluntary guidelines for the commercial spyware industry at the Paris Peace Forum this week, marking the latest step in an ongoing international effort to curb the proliferation of surveillance tools whose reach is being amplified by advances in artificial intelligence.
The development matters because it represents one of the only coordinated multilateral responses so far to a surveillance technology market that researchers say is expanding faster than governments can regulate it, with AI increasingly cited as a factor lowering the technical barriers for both state and non-state actors seeking to deploy intrusive tools.
According to the Stimson Center's Allison Pytlak, who attended the Paris proceedings, the new Industry Guidelines were developed through the French- and UK-led Pall Mall Process and are structured around four pillars: compliance with regulations and norms, due diligence, risk management and monitoring of compliance. They build on a 2024 voluntary Code of Practice for States, which is organized around accountability, precision, oversight and transparency and is currently endorsed by 27 countries.
One unnamed speaker at the conference described the Pall Mall Process as a
wrapper
that consolidates previously fragmented efforts to address what has been called one of the murkiest markets in existence, referring to the commercially available cyber intrusion capabilities industry that sits at the intersection of legitimate security research and offensive surveillance.
Citizen Lab's recent research into a newly documented form of zero-click spyware delivered through targeted advertising, dubbed
AdFection
by researchers, was referenced repeatedly during the Paris conference as evidence of how rapidly the threat is evolving. Conference participants also cited Apple's new feature that sends push notifications to users suspected of being targeted by mercenary spyware, which according to reporting cited by Pytlak has already triggered an estimated wave of alerts across 99 countries, reaching legislators and government employees in Germany as well as journalists in Turkey.
Pytlak noted that civil society organizations including Access Now, Amnesty Tech and Citizen Lab have been described by one conference speaker as the
motor
driving years of documentation and advocacy on spyware harms, and argued that the evidence base is now sufficient that the burden has shifted to governments and industry to act on it rather than requiring further proof of harm.
A separate technology assessment released this month by the US Government Accountability Office found that AI technologies have already magnified the reach of traditional cyberattacks, enabling high-quality phishing content, deepfakes and more precise attacks, while also noting the emergence of agentic AI systems capable of autonomously carrying out multiple stages of an attack with minimal human input. The GAO identified a June 2026 case in which a threat actor used multiple AI agents to test malicious code against security products and map new attack techniques to known frameworks.
Industry researchers at LevelBlue separately documented a broader pattern of cyber operations increasingly extending beyond military and government targets into utilities, telecommunications and critical infrastructure, including compromises of emergency-warning applications in Ukraine and Russia in November 2026 that distributed false nuclear and chemical threat alerts, illustrating how such attacks can undermine public trust without causing direct physical damage.
Whether the newly adopted voluntary guidelines will meaningfully constrain the spyware industry remains an open question, with conference participants themselves debating whether firms will comply absent binding legal consequences, even as the diffusion of AI-enabled tools continues to expand the range of actors capable of deploying sophisticated surveillance and intrusion capabilities.