As a public servant, I increasingly see the potential of artificial intelligence tools to improve efficiency in our day-to-day work — whether by generating concise summaries of lengthy documents, or by helping us make sense of complex information more quickly. At the same time, I am acutely aware of the risks, particularly when it comes to handling sensitive or protected documents, where inadvertent disclosure or reliance on non-secure platforms could undermine the very principles of confidentiality and trust that underpin our work. Read MoreA former Chief Human Resource Officer explains the ‘prudent path forward’ when it comes to using AI in the federal government.
A former Chief Human Resource Officer explains the ‘prudent path forward’ when it comes to using AI in the federal government.
Article content
 |
Dear Public Service Confidential,
As a public servant, I increasingly see the potential of artificial intelligence tools to improve efficiency in our day-to-day work — whether by generating concise summaries of lengthy documents, or by helping us make sense of complex information more quickly. At the same time, I am acutely aware of the risks, particularly when it comes to handling sensitive or protected documents, where inadvertent disclosure or reliance on non-secure platforms could undermine the very principles of confidentiality and trust that underpin our work. Given this tension, I find myself wondering: what guidance can public servants rely on when it comes to responsibly using AI in the workplace? Should there not be a government-wide policy, rather than the current patchwork where some departments permit the use of tools like ChatGPT while others prohibit them entirely? Without clarity and consistency, employees are left in a difficult position — caught between the promise of innovation and the fear of overstepping unseen boundaries. What, in your view, is the prudent path forward? — a public servant exploring responsible AI use. |
Article content
Story continues below
Article content
Article content
Article content
 |
The answers to your questions will probably do more to shape Canada’s future public service and institutions than most budget reduction exercises. That’s why thinking about them now is critically important.
I experienced the arrival of desktop computers, word processing and internet access in government workplaces (also faxes and the end of in-office smoking, but those are separate stories). Each arrived with breathless descriptions about how they would change the workplace and government services beyond recognition. Computers would produce quantum productivity gains. Self-service word processing would eliminate thousands of jobs. Internet access meant that we would always know what Canadians wanted to tell us. We have definitely experienced real benefits, but they often differed significantly from the original narrative. To be clear, artificial intelligence IS coming, en grand, and soon. Before long, it will be impossible to remember having operated without it. It must be remembered, though, that artificial intelligence is only a tool — a particularly powerful one with important qualities that differ from most others, but still just a tool. As with all tools, its ultimate impact will be significantly shaped by the choices about where, how, and why it is used and with what safeguards. Deliberate choices need to be made about the data we allow to be used in or excluded from artificial intelligence use and processing. Government leaders will need to make decisions about the safeguards needed to protect personal, sensitive and secret data. We need to understand any ownership, control, access and possession implications once this data is retrieved and processed by federal government or external parties’ artificial intelligence capabilities (something that many, including some Indigenous governments, have expressed concerns about). Important judgments need to be made about identifying and addressing the biases that can be inherent in artificial intelligence decision-making and products. It would be best to have a clear understanding of the proposed responses to cases of significant failures before they need to be mobilized for the first time. The accountabilities for having these safeguards in place and for ensuring that they are respected must be clear, transparent and sufficiently resourced. As with all new technologies, we can expect to encounter, over time, critical questions that we cannot even imagine today. It’s hard to envision an effective whole-of-government response to those needs if core guiding principles vary from department to department. It requires a single government-wide set of core policies and standards that provide an agreed upon range of flexibilities. We once decided that having two separate word processing packages presented more challenges than we cared to manage (the great WordPerfect vs Word debate of the 1990s). Will having multiple artificial intelligence platforms prove any different? The way that public institutions will be using artificial intelligence 20 years from now will differ greatly from today’s promises and speculation. History suggests that much of that change will come from unwanted, accidental and avoidable setbacks and learning. Your question suggests that we make as much of that learning as planned, deliberate, coherent and value-driven as possible. I think that you are entirely right. — Daniel Quan-Watson, Public Service Confidential |
Article content
Story continues below
Article content
Daniel Quan-Watson was a deputy minister and CEO with the federal public service for nearly 15 years. He was also a provincial public servant with the governments of Saskatchewan and British Columbia in the early stages of his career. He is a pilot and rode the Dempster Highway to Inuvik on a motorcycle.
Article content
Article content
Read More
-
Do public servants need to be afraid of artificial intelligence?
-
Four things public servants need to know about the federal government’s new AI strategy
-
Advertisement embed-more-topic
Story continues below
Article content
Are you a public servant with questions about your workplace? Write to us anonymously at PSConfidential@postmedia.com and we’ll pick our favourites to send to an expert columnist. No gripe is too small. No topic is too big.
Article content
Public Service Confidential is an advice column, written for the Ottawa Citizen by guest contributors Scott Taymun, Yazmine Laroche, Daniel Quan-Watson, V. C. de la Ronde and Chris Aylward. The information provided in this series is not legal advice and should not be construed as legal advice.
Article content
Article content