In a recent report published in June about the integration of unmanned aerial vehicles (UAVs), or drones, into the U.S. national airspace, the National Academies of Science, Engineering and Medicine argue that the Federal Aviation Administration (FAA) should develop new ways of assessing risks associated with unmanned aircraft operations.
Basically, the report “Assessing the Risks of Integrating UAS into the National Airspace System” states that the FAA is too risk-aversive, thereby forfeiting many social advantages that drones potentially have. Hans Herkeens – assistant professor at the University of Twente, Netherlands, and chairman of the Platform Unmanned Cargo Aircraft (PUCA) – here discusses a few salient takeaways of the report. He argues that the recommendations in the report ignore some practical roadblocks to a more risk-permissive approach to safety, but provide pointers for facilitating the introduction of drones.
Gross or net safety?
According to the report, the risks of drones as assessed by the FAA are overstated. An important reason is that the potential of drones to enhance safety is ignored. For example: having drones instead of people inspect cell phone towers will make this work far less hazardous. In my opinion, this reasoning is sound in theory, but it ignores the practical problem that accidents that are caused by new artefacts like drones are likely to cause more public unrest than accidents caused by applications that people are familiar with. Look at the controversy incidents with self-driving cars generate. This is especially true when costs and benefits fall to different groups, for example the general public who fears drone accidents versus workers whose occupation is made safer. Besides, accidents that are prevented are conspicuously less visible than accidents that do happen.
Perceptions and facts: how safe is safe?
A similar practical obstacle can be identified concerning the recommendation in the report that the FAA assess the risk of drones that the public will accept in the same context as other risks like traveling by car or walking across the street. The implication is that drones pose no significant risk. But the perception of people may not only be different from actual fact when asked about drone accident risks, but it may also change dramatically when a serious accident occurs that people feel, correctly or otherwise, could also happen to them. In decision making research there is ample evidence that people often do not perceive their preferences (risk tolerance) and actual risks correctly.
Furthermore, they perceive losses as more harmful than equivalent gains. So, if drones save ten lives but cause the death of ten people, the net effect is zero but is likely to be perceived as negative. Also, the average person will likely perceive the risk of drone accidents as a net risk increase on top of existing risks. And we should not forget that, in the coming years, the number of drones flying may well be a factor of thousand over that of aircraft. Even if the chances of an accident per flying hour are the same for drones and aircraft, there may still be a considerable number of drone accidents. The overwhelming majority of them will not be fatal, but even non-fatal accidents may seem threatening when occurring often enough. Of course, drones give other benefits than safety: the possibility to have goods delivered quickly to one’s doorstep, the fun of flying recreational drones. It takes, however, special care to balance this with safety risks, because research seems to indicate that people have great difficulty in comparing apples with oranges. This is probably why people are often interested in, and concerned with, aircraft crashes, but in the end usually choose a flight based on ticket price.
The practical obstacles discussed above do detract nothing from the report’s value. Reducing overly conservative safety demands and balancing safety risks with benefits are possible and needed, but a well thought-out drone introduction strategy is paramount. This strategy should address the following elements:
1: Pro-active communication with politicians, policymakers, opinion leaders and the general public. Explain the benefits, but also be open about risks before, not after, the first accidents happen, as they inevitably will. Research indicates that people generally will accept drones for, for example, transport, when properly educated about them and when they are aware of examples of useful applications.
2: These useful applications should be actively pursued, not waited for. Projects for, for example, disaster relief should be embraced and promoted.
3: The introduction of drones should be phased. This is perhaps difficult with small recreational drones, of which so many are already flying, but it should be possible with small delivery and disaster relief drones and with unmanned cargo aircraft (UCA). Operate them initially in areas where both the chances and consequences of accidents are small, and only then start operations in, for example, densely populated areas. It would be best to conduct a number of increasingly complex real-life case studies and evaluate them systematically. Such a phased approach could also entail a gradual relaxing of safety requirements along the lines of the National Academies report. Acquiring experience in a structured way would also partially solve a problem identified in the report: the lack of statistical evidence of the safety of drones. Hopefully, by the time the first serious accidents occur, drones will be so commonplace and the benefits so tangible that the reaction of the general public will be largely rational. By that time drones should be so safe that a rational reaction is, well, rational.
The National Academies report gives a refreshing look at drone safety. The observations made seem correct to me – in theory. But without a careful, phased, drone introduction strategy, practical obstacles could stand in the way of fully reaping the benefits of drones. Governments should act pro-actively, not only in the U.S. Presently, in many countries the emphasis seems to be on regulations aimed at mitigating risks, whereas it should be on policy aimed at maximizing benefits, with an acceptable level of risk.
Republished with permission from ADIT – The Bulletin.