I visited the Kennedy Space Center in 2010, where a Saturn V moon rocket was exhibited with an explanation that the computer power contained in it was equivalent to that of a 2010 desktop computer. Since that visit, computer power has increased 100-million-fold, leading to the advances in artificial intelligence (AI) that is a significant cause of concern at societal level.

Earlier this year, the BMJ Global Health published a paper by Federspiel et al¹ on threats by AI to human health and human existence, cautioning on the potential negative impacts on humanity of this exponential growth in computing power. The potential of AI in diagnostics, bio-informatics and cancer research is significant but needs to be counterbalanced by its potential threat to democracy, peace and safety, and work and livelihoods. Many authors have highlighted the significant investment that is occurring in this area and that the pace of regulation and oversight lags this significantly. 

An added threat is the potential weaponisation of AI in the cyber environment. A previous phishing cyberattack on Ireland’s national health system in May 2021 has had a persistent legacy to this day due to theft of patient and staff records. But what if a cyberattack also corrupted records including personal data and healthcare records? Equally, in scientific research, manipulation of data through algorithm manipulation would be much harder to detect than traditional scientific misconduct detected by audit, in the case of breast cancer surgical and high-dose chemotherapy trials. 

The article by Federspiel et al is a cautionary clarion call for healthcare in this regard and hopefully will prompt a reflective review of this area by regulators, legislators and researchers.

Reference

1. Federspiel et al. BMJ Global Health 2023; 8: e010435