The rise of autonomous vehicles (AVs) and drones is transforming transportation and logistics. From robotaxis to last-mile drone deliveries, these developing technologies promise lower costs, greater efficiency, and potentially safer roads.

But how safe are they, really and what are the safety concerns surrounding these innovations and their potential impact?

The potential of Robotaxis lies in the aim that they redefine urban transport by removing the need for human drivers. By cutting costs to an estimated £0.20 per mile, these autonomous Robotaxis could make personal car ownership less appealing, boosting the taxi market to an estimated £4 trillion. However, as always, their adoption hinges on public confidence in their safety.

So what are their safety records so far?

Companies like Waymo report over 20 million miles of fully autonomous driving, with significantly fewer accidents compared to human-driven vehicles. It seems that human error such as fatigue, texting, or otherwise impaired driving is responsible for the vast majority of car crashes. Clearly, autonomous systems eliminate these risks.

However, scepticism remains and despite the statistics, experts caution against overconfidence as the 70 million autonomous miles cited by industry advocates often included a human safety driver ready to intervene.

Philip Koopman, from Carnegie Mellon University, Pittsburgh, US, argues that true safety requires at least 300 million miles of driverless testing with zero fatalities, but what makes autonomous vehicles safer?

Several advancements underpin the safety of autonomous systems:

Firstly, machine vision means that technologies like LIDAR and cameras help AVs detect obstacles with high accuracy.

Secondly, AI and Machine Learning allows vehicles to navigate roads, interpret traffic patterns, and react to unexpected scenarios.

Thirdly, incremental rollouts by companies prioritise safety by rolling out autonomous features in phases, often with human oversight.

However, limitations remain. Autonomous systems lack “common sense” and can fail when encountering unfamiliar situations. Because these systems rely on training rather than intuitive problem-solving, it can lead to unpredictable behaviour in rare scenarios.

There is also the issue of regulation and public trust. Unlike conventional vehicles, autonomous systems currently lack uniform safety regulations. Instead, individual states or countries set their own rules. This inconsistency inevitably complicates large-scale deployment and consumer confidence.

So what needs to change?

The development and implementation of a set of unified standards could mean that clear, enforceable regulations could ensure uniform safety benchmarks, while more transparency by companies disclosing more data on their testing – for example, the inclusion of miles driven without human intervention.

In addition to the potential use of autonomous vehicles as taxis, deliveries using drones is also on the rise.

Drones are increasingly facilitating logistics, particularly for last-mile deliveries, and their ability to navigate traffic-free airspace could reduce costs by 57% to 94% depending on the package and delivery circumstances.

Thers is also the life-saving potential of drones making an impact in critical situations. In Rwanda, when drones were able to deliver blood for transfusions in under 15 minutes, post-partum haemorrhages were reduced by 80%.

Nevertheless there are challenges ahead.

Understandably, regulatory hurdles remain a significant barrier, particularly for flying drones in densely populated areas, and while ground-based delivery robots are further along in testing, there are still safety concerns about their interaction with pedestrians and traffic.

Ultimately, autonomous vehicles and drones hold immense potential to reshape transportation and logistics. Robotaxis could make commuting cheaper and reduce accidents caused by human error, while drones could revolutionize delivery services and save lives in emergencies. Yet, safety remains a critical concern.

The industry must continue rigorous testing, improve AI’s decision-making capabilities, and adopt uniform safety regulations. Until then, cautious optimism is warranted as we navigate this transformative shift.