In this paper, we argue that to accommodate the change brought by
autonomous and semi-autonomous cars, the current lifecycle model of a
car should be re-assessed. Software security issues for vehicles are
amplified significantly, as vehicles become more dependent on software
and the driving process is transitioning towards autonomous operation.
As vehicle automation levels increase and fully autonomous cars are
already on public roads, the options for malicious parties to influence
and harm traffic by hacking vehicles multiply, increasing risk levels of
traffic. The magnitude of this issue is huge as most countries in
Western world are dependent on the road travel, and the problem is a
multidisciplinary one ranging from business, technology, and security to
juridical, sociological, and ethical issues. We state that as the
digitalisation and AI-based solutions take part to driving processes, we
must start thinking cars more as a service and not as a standalone
product. We argue that there should be more focus and further research
on the issues of software maintenance and new software models to keep
the roads safe from outdated and dangerous automated or semi-automated
vehicles.