NHS restricts open-source code amid AI security fears, sparking debate over transparency vs. protection in public health infrastructure
Original framing: “Backlash builds over NHS plan to hide source code from AI hacking risk” — New Scientist
Peer-reviewed studies in *Nature Digital Medicine* (2022) show that open-source medical algorithms have lower error rates due to crowd-sourced debugging, contradicting the 'security through obscurity' myth. The NHS’s reliance on closed-source AI models (e.g., Mythos) introduces vendor lock-in, a documented risk in healthcare IT that increases costs and reduces adaptability. Cybersecurity research consistently demonstrates that transparency is a stronger deterrent to attacks than secrecy, as seen in the Linux kernel’s resilience.
The NHS’s decision to restrict open-source code reflects a broader neoliberal trend in public health, where 'security' is weaponised to justify proprietary control, obscuring the fact that open systems have historically been more resilient.
This move aligns with the interests of AI vendors like Mythos, whose backers profit from the illusion of vulnerability, while marginalising the voices of clinicians, patients, and Global South partners who depend on transparency. Indigenous and cross-cultural models—from Māori *kaitiakitanga* to Brazilian *SUS* mandates—offer a radical alternative: health data as a communal resource, not a corporate asset. The solution lies in federated, participatory systems that distribute power as much as they distribute risk, ensuring that 'security' serves the public rather than the powerful. Without this shift, the NHS risks repeating the failures of proprietary healthcare IT, where opacity breeds both cybersecurity threats and democratic deficits.