The Future of Federal Safety Oversight, originally published at CorregidorTech.com.
— — — — — — — —
Dear Honorable Pete Buttigieg
The subject of my writing today is to introduce the life-saving opportunity that the leader and future leaders of the Department of Transportation, and other government safety oversight organizations, have before them, provided by the unique benefits of technology and, especially, modern data analysis.
I do not write today in my professional compacity, in which I help develop digital applications for the Department, or to suggest that my opinion is shared by those with whom I worked while a civil servant in the Department of Transportation for the better part of a decade. I write today to communicate observations borne of those experiences, and the associated lessons learned. In short, these opinions are solely my own.
Using the Federal Aviation Administration — the mode with which I am most personally familiar — and specifically its aircraft design and manufacturing oversight responsibilities, as an example, we may conclude that the agency’s oversight methodology has been effective resulting in “the chances of perishing in an airliner [being] lower than ever”. From my experience, that oversight process, however, has demonstrated to be mostly retroactive in nature; that is, the agency provides regulatory guidance in response to safety events and schedules the frequency and targets of its oversight activities (audits, inspections) based on the results of a regulated organization’s previous year of quality and compliance history.
Recognizing that an even higher standard of safety may be achieved by more forward-looking oversight practices, the regulatory organizations have adopted the principles of Safety Management Systems (SMS) and, related, a renewed relationship paradigm between the regulated and regulator, applying the “Accountability Framework” and encouraging voluntary disclosure by regulated entities of noncompliances in design and manufacturing processes. These safety oversight process improvements should enable the FAA to apply risk-based resource targeting and data-driven decision-making when determining whether regulatory action is needed to correct a safety event related to a design and/or manufacturing quality escape, and when determining where and how often to audit and inspect the design and quality systems that the FAA has the responsibility to oversee.
Adopting these values was a necessary first step by the FAA in the upward climb toward progressively effective safety oversight. But the adoption of these values is not, in my observation, being leveraged to its full potential. When these values are fully internalized by the FAA and, hopefully, across a wider spectrum of safety oversight organizations, the evidence will be demonstrated in refreshed processes that are not reactive but proactive in nature.
What does that outcome look like specifically? A regulatory environment in which the federal oversight personnel has at his or her fingertips data insights that describe the regulated entity at both an overall, corporate level, and targeted to the individual facility, supplier, and production line, allowing the regulator to make real-time oversight decisions based on the understanding of the regulated organization’s values as a whole, and its performance at the ground-level regarding the design and production of the parts and products that the regulator is individually tasked to oversee.
There are many ways to harness technology to shift oversight practices from being performed in response to negative events and, instead, being enacted to mitigate future occurrence of undesired outcomes. It is not one person’s purview to describe them all, and certainly beyond my own aptitude, but some immediately apparent examples are provided below. The intent here is not to ideate solutions, but to compel the leadership spirit that may embrace remedies such as these, or those like these in their pursuit of supporting proactive, data-driven safety oversight decision-making:
· Open-sourcing internal FAA data analysis tool sets so that regulated organizations may inherit and integrate them into their own technologies thereby encouraging the real-time sharing of data between regulated and regulator, resulting in both more effective targeted oversight and less oversight activity that is unneeded, ineffective, and potentially costly to the regulated entity;
· A mode — or agency — wide data consolidation effort that collects disparate data including service difficulty reports, compliance and enforcement records, quality management system data, operator reports, voluntary disclosure information, et. al, that can help the regulator better understand the overall organizational health of the entities he or she oversees across all the CFR Parts — and visualize trends, via machine learning, in seemingly minor quality degradations that may ultimately represent a decline toward a future major safety occurrence and, essentially, provide the opportunity for the safety oversight authority to assist the regulated entity in mitigating it;
· Better management of designees and delegated organizations by making the type of consolidated data described above available to them and their business intelligence tools to empower the industry experts to better assist the federal agency on whose behalf they serve. Promote increased, barrier-less communication between federal agencies and their designees and delegated organizations via modern technology communication media, with the added benefit of real-time data about designee performance being made available to those within the federal agencies that have the authority to limit or expand delegated responsibilities;
· Harnessing non-traditional data storage and data analysis techniques including non-relational, big-data persistence via cloud service providers and cutting-edge business intelligence and data visualization tools to augment existing safety oversight continued operational safety processes. For a specific example, grow the Monitor Safety / Analyze Data (MSAD) process implemented by the FAA Aircraft Certification Service (AIR) as an SMS tool by feeding it the standardized, voluminous data that it needs to be effective, and strengthening it by utilizing the libraries of today’s data science systems in order to perform rigorous risk analysis and produce effective predictive models.
Once these systems are adopted, they may be integrated with similar automated systems developed by allied civil aviation authorities, potentially supported by ICAO, truly unlocking global safety and quality observations, allowing safety oversight authorities to perceive the holistic, overall wellbeing of the multinational organizations they have the responsibility to regulate.
These are the “low hanging fruit” that may be immediately or nearly immediately adopted to apply the same technology principles that are “disrupting” ubiquitously other industries and revolutionize the safety oversight model. With the strategic objective being the successful shift from reactive to proactive processes.
To some degree, agencies are already striving toward the adoption of these, or similar, systems. But, from my view, the implementation has been haphazard, presenting the need, as I am describing here, for an influential voice to unify and drive forward this vision and to make data-driven decision-making not just a buzzy catch-phrase, but an internalized organization value that all members of the safety environment embody, work to adopt, and embrace when provided its tools.
New leadership provides refreshed opportunity for the bold to embrace change that is significant — and difficult — but also potentially lifesaving and paradigm-shifting in its effect.