The Paradox of Innovation and Continuity
Public institutions live in a perpetual state of creative tension: they must innovate without losing the very traditions that make them trustworthy. As Christopher Pollitt (2011) reminds us, reform in public administration is rarely a clean break with the past; rather, it is a careful balancing act between novelty and the enduring norms of accountability and transparency that give our work legitimacy. For those of us working in parliamentary reporting, that paradox is not a theoretical one. It plays out daily in the push and pull between the speed of modern technology and the deliberative rhythm of the parliamentary cycle.
At the Legislative Assembly of British Columbia, Hansard Services confronted that tension head-on when we developed Parrot — our in-house automatic speech recognition (ASR) tool. We could have outsourced the work to commercial vendors, but we chose another route: one that protected our independence, respected our editors’ expertise, and anchored innovation in purpose. The story of Parrot is less about technology than about professional identity, public trust and the quiet conviction that modernization need not mean mechanization.
Innovation, Public Value and Professional Identity
Public administration scholars have long emphasized that innovation must serve public value rather than novelty for its own sake. Mark Moore (1995) defined public value as the alignment of efficiency with legitimacy the union of effective service and civic trust. In that spirit, Mergel, Edelmann and Haug (2019) argue that digital transformation requires not only technical upgrades but cultural change, collaboration and citizen-centred design. For Hansard, the question was not how to automate transcription, but how to strengthen the human judgment that gives a transcript its meaning.
Sørensen and Torfing (2017) describe sustainable change as the capacity to balance innovation with continuity – to foster trust while enabling new ways of working. Leadership, in this view, is not about disruption but about accompaniment. In Hansard’s case, that meant treating editors not as subjects of innovation but as co-authors of it.
Hansard in British Columbia
Hansard Services produces the official verbatim record of legislative debate and committee proceedings. Increasing workloads, extended sittings and growing public demand for immediacy created both strain and opportunity. Commercial ASR products beckoned, but at a price: dependency, cost escalation and potential loss of control over our data – and by extension, our reliability and trust.
So we built our own. Parrot emerged from two guiding ambitions: first, to sustain internal capacity and reduce reliance on external vendors; and second, to embed technological innovation within our existing culture of professionalism. Based on OpenAI’s Whisper model, Parrot processes audio directly from the Chamber, generates draft transcripts, enriches them with metadata and hands them to human editors for refinement. The machine provides the momentum; the humans, the meaning. As one editor observed, Parrot doesn’t understand the words it repeats – but it does give us a head start in understanding one another.
A Participatory Change Process
Change imposed is change resisted. We knew that success depended not on persuasion but participation and ownership. So rather than announce a shiny new tool and wait for applause, we began with small focus groups, open discussions and a healthy dose of skepticism. Editors and developers worked side by side – one committed to subtlety and nuance, the other understanding how to convert the over-arching principles into code – in a process that resembled dialogue more than deployment.
Our team ranges in age from their 20s to their 70s, a remarkable spread of professional experience and technological comfort. That diversity proved to be an advantage. At the risk of over-generalizing, younger colleagues brought curiosity and daring; senior editors brought the discipline of craft and the institutional memory that kept us grounded. Training was shaped around both enthusiasm and hesitation, acknowledging that progress moves at the speed of trust.
There were anxieties, of course. Some wondered whether an application that could transcribe so quickly might one day render editors redundant. We addressed those concerns directly: no jobs lost, no craft diminished. The goal was to save time, not replace talent. Innovation, after all, is rarely about replacing the human mind; it’s about giving it a little breathing room to think.
Cultural and Operational Transformation
The results were tangible. Average editing time for a five-minute take dropped from 45 to 35 minutes – a 22 per cent time saving. Overtime costs fell; fatigue lessened. More importantly, the perceptions changed. Editors began to see Parrot not as a rival but as a collaborator – a particularly diligent amanuensis, fast and tireless, occasionally erratic, and always in need of a discerning editor.
By keeping development internal, Hansard preserved its autonomy and data security, avoiding the vendor lock-in that can quietly erode public accountability. Yet the deeper transformation was cultural. Those who might have found themselves comfortable in the “resister” camp began to think differently about innovation: not as disruption, but as stewardship.
Ethical AI and the Human Dimension
The Parrot project also reframed how we think about the relationship between human expertise and artificial intelligence. Rather than treat AI as an existential threat, we approached it as a partner in professional work. Bullock (2019) argues that legitimacy in AI governance depends on transparency, accountability and human oversight – precisely the qualities Hansard sought to preserve.
By embedding AI within our existing editorial ethics rather than re-engineering those ethics around the machine, we maintained continuity of professional standards. Editors retained final judgment; the record remained authentically human. As Sørensen and Torfing (2017) put it, technology may be the catalyst for change, but people provide its meaning. In practice, that meant open dialogue, shared learning and the recognition that empathy is as essential to innovation as engineering.
Innovation as Stewardship
The story of Parrot suggests that modernization, when rooted in public purpose, strengthens rather than weakens democratic institutions. Through co-creation, transparency and respect for professional identity, Hansard integrated AI in a way that honoured both the craft and the calling of parliamentary reporting. Pollitt (2011) reminds us that sustainable reform depends less on clever tools than on thoughtful continuity. Our experience affirms that view.
When technology is guided by human purpose – when it gives the mind a little more breathing room to think – innovation becomes an act of stewardship. Parrot does not replace human editors; it amplifies them. In doing so, it reinforces the quiet truth that progress and tradition, like two consecutive pages of a well-edited transcript, make the most sense when read together.
D’Arcy McPherson is the director of Hansard Services at the Legislative Assembly of British Columbia, Canada.
References
Bullock, J. B. (2019). Artificial intelligence, bureaucratic form, and discretion in public service. Information Polity, 24(4), 451–466.
Mergel, I., Edelmann, N., & Haug, N. (2019). Defining digital transformation: Results from expert interviews. Government Information Quarterly, 36(4), 101385.
Pollitt, C. (2011). Innovation and public sector reform: Concepts, themes and paradoxes. Public Administration, 89(1), 1–20.
Sørensen, E., & Torfing, J. (2017). Sustainable change in public administration: Balancing innovation with continuity. International Public Management Journal, 20(3), 389–405.
