Blog: Week 9 — A Week of Learning
Reflections on a week spent largely out of the office learning
Another week has passed and it is time to share what I’ve been working on. I’m surprising myself that I am still working on the weeknotes 9 weeks in. I didn’t think I would last 3 to be honest. However, publishing weeknotes has allowed me to make connections across the public service so beyond the useful practice organizing and sharing my thoughts, blogging has allowed for interesting and meaningful connections to flourish. For anyone on the fence, I highly recommend you try talking about what you’re working on and see what happens. It is a commitment to write up my notes every week but weeknotes are not a formal blog post so I find I’m ok writing my thoughts more free flow than structured much like my own mind likes free flowing ideas than structure.
This week, Abe Greenspoon, Amanda Bernardo, Claudia Levac, Aalya Dhanani Essa and myself hosted the first ever OneTeam Gov Can lunch time meet-up. Why start a lunch time meeting? I was finding it hard to attend the existing meet-up (8:30AM every second Wednesday) so I connected with other like minded public servants to start a lunch time meet-up hopefully opening up the beauty of a shared network and inspiring conversation to a wider audience. The plan is to host lunch time sessions on Thursday every 2 weeks with alternating sessions (starting in June) being 100% virtual and open to public servants at any level of government from across Canada.
This week, the Auditor General released a new report on the number of calls dropped at the larger call centres in the Government of Canada (CRA, Veteran Affairs etc.) The report found that nearly 50% of the 16 million calls Canadians made in 2017–18 were dropped or otherwise not answered. This is a staggering finding and most likely reflects your past lived experience with a government call centre (it definitely reflects mine). Among the many findings in the Auditor General reported included the lack of service standards for call centre wait times, aging systems, and that many of the Departments examined did not assess the needs of their clients before making changes to their call centres. Sound familiar? This is not a new problem. For as much progress as call centres have made, there are still chronic problems that plague the system.
In a past life, I worked on the Service Policy file. This included having a pulse on what was happening with call centres, understanding how they worked and establishing service standards for everything from wait times, call lengths etc. Throughout my work (and subsequent work where I got to shadow call centre employees) one thing stood out. The tireless and (often) thankless hard work done by front line call centre workers. They are the face (or more accurately voice) of the Government of Canada and for the most part we couldn’t ask better people to do the job. However, many Departments have chronically underfunded their call centres. On top of that, the push for digital first service delivery means many are losing sight of the non-digital infrastructure needed to ensure fair and equitable access to government services.
What do I mean by fair and equitable access to government services? Regardless of where you are in Canada and what level of access to modern high speed (broadband) internet you have (and in many cases that you don’t have) the Government of Canada has a moral obligation (“it must”) provide access to its services. Call centres become an integral part of service delivery. Even if there is a robust and mature digital service delivery model in place, we must create with omni-channel delivery in mind meaning seamless transition between digital, phone and in person. Failure to build seamless transition means we are creating unfair and potentially discriminatory service delivery.
Future Forum and UX in AI/Machine Learning
As the title eludes to, I spent much of this week out of the office at events learning about digital government, AI, machine learning and other topics of how the world is changing and what that means for both government and governance. Beyond the immediate relevance to my existing work there was the opportunity to learn and reflect on how the world was changing and what that would mean for the way we run and govern Canada.
On Monday and Tuesday, I attended the Institute on Governance’s Future Forum, a 2 day event focusing on top of mind digital issues that governments will soon face. An excellent set of speakers highlighted a lot of interesting information including addressing the potential for bias in AI (more on that later), policy implications of relying on AI for decision making and busting a few myths about new and emerging technologies. While I found the conference insightful I noticed a gap in the room in terms of technical expertise, level of discussion and how speakers presented their topics. Many of the talks and issues brought up focused on things that were happening right now. For a conference branded “Future Forum” most of the speakers were not talking about the future and the disruption by advances not already happening. Perhaps this was a reflection of the interests of the audience or that government is perceived as already behind the curve.
The other conference I attended was the IoT613 UX workshop for AI/Machine Learning. A day long workshop focused on diving deep into how AI/Machine Learning applications are best developed to fill a gap or solve a problem that existing users have. What I found interesting about the workshop was the focus on normalizing AI/Machine Learning, showing interesting examples of this technology in the wild and in a way taking away the magic to show that the AI we are making is only as smart as we make it to be.
Bias in AI
More a teaser for this week as I realize how long my weeknotes are becoming with no end in sight.
Over the past week, the issue of bias in AI keeps coming up in talks, presentations, tweets and discussions. There is a wrongly held assumption that data is unbiased and thus AI itself (driven by data) will be unbiased. Of course we know this isn’t true. AI (and the data we use to train AI) is as bias as we are. Racialized individuals (especially women of colour and Indigenous people) face systemic discrimination and other barriers. Whether we see it or not, the playing field is not level. As a result, the data we generate has those same bias’s since data is a reflection of the instituions or individuals who generated it. AI is not a silver bullet. It will not suddenly create a system of merit where the most qualified person rises to the top. Why not? Because society itself is not merit based. Because workplace hires are not determined by merit (no matter what anyone says). Until we get rid of outright and unconscious discrimination, we need to recognize that the data we are using to build our AI systems and platforms is bias. There are steps we can take to reduce bias (a good place to start is recognizing the bias our existing systems already have and then enacting policies and other fixes to eliminate bias). However, the notion that AI itself leads to a system free of bias is wishful thinking at best and laughable nonetheless.
AI Project Update
A short update on both AI projects this week.
The Regulatory Evaluation Platform is still live for bidding. Bids are due on May 22 and for obvious reasons I will not be able to discuss the process or bids until we award up to 3 companies contracts to build prototypes. Don’t expect much news on this project for at least another month.
The Incorporation by Reference project hit an exciting milestone this week. After many weeks of hard work and consultations with regulators, we have started the AI training process. This is being done by enlisting regulators to help train the AI on what is an incorporation by reference as well as some other details. With enough training data sets completed, the AI will be able to recognize a real Incorporation by Reference by itself with at least 80% certainty.
What’s on my mind?
What’s on my mind? Get in touch if you want to chat or have an insight to share.
- Why is the culture of the public service not changing? More specifically, why is the culture changing in pockets incrementally but many more still face work environments which reflect a past era of management?
- Year over year, trust in the public service declines. What can the public service do to restore Canadians trust?
- With the eventual collapse of society due to climate change, does public sector innovation (and the struggles many public servants go through) matter?
Week 9 of my weeknotes are in the books. Thank you for reading and hope you have an awesome week! See you in 7 days!