AI and Public Services

Today, I read two articles that, at first, appear to be entirely unrelated to one another.

Technology – Another Drain on Public Resources?

First, this one:
Elizabeth Thompson, “Federal Government Plans to Increase its use of AI – With Some Big Exceptions” (CBC News, May 27, 2024), online.

I read this article with interest. I have a continuing concern about the persistent underfunding of our public services in this country.

I’ve seen it happen first hand, and have written about here. Technology companies come to the public sector looking to make millions. Technology companies step up to the plate, make their bid, get the contract, build the thing, and then leave. Ultimately, due to a lack of public sector resources to provide the much-needed oversight of these technologies, trouble ensues for the people on the ground who are trying to do their jobs with their hands tied behind their back for years afterward.

Now flashforward to the federal government’s new announcement that it is going to pay out $2.4 billion dollars to “secure Canada’s AI advantage”. Presumably the lion’s share of this money will be handed over to corporations with few strings attached, chasing the spectre of AI economic advantage. However, as Kean Birch wrote on April 4, 2024 for the Globe and Mail, “Generative AI is Simply a Waste of Time and Money”.

I tend to agree. I’d rather have my tax dollars used to reduce emergency room wait-times, thank you very much.

Or a Useful Tool?

Later in the afternoon, this article crossed my desk:

Dean Beeby, “80-Year Extension on Access-to-Information Request Appears to be a Record” (CBC News, April 13, 2018), online.

The article goes on to detail a 70 year old citizen’s freedom of information request that was made to Library and Archives Canada (LAC). He was looking for the results of an RCMP probe from 1993 on money laundering and corruption (which resulted in no charges being laid).

The requestor later received a response from LAC indicating that they will be unable to respond for 80 years.

Really. Shocking. But understandable – there were 780,000 document pages in question, requiring review and approvals from different departments.

So, then my mind turned to some decent e-discovery tools, available or in development in the legal space that purports to review large swaths of documents in a very short span of time. It makes me wonder whether any government institution that is required to respond to FOI requests – should be invested in this technology to speed up servicing these requests?

Ultimately, I have questions, like:

1. How well does this technology work for digitized content? Having been involved in digitization projects in the past, digitizing print documents can take a significant amount of time based on the volume;
2. How is the quality of a digitized record actually ensured so as to ensure a quality OCR output, sufficient for reading by (anyone), including AI;
3. How can information professionals (everyone involved in records management for government) get up to speed quick on how to acquire and implement this type of technology;
4. What are the other plausible (meaning useful, not a money pit, not terribly disruptive) uses of AI to improve government services?
5. Is any of the government’s 2.4 billion going to be used to improve government services? It doesn’t look like it?

Welcome to the jungle.

The post AI and Public Services appeared first on Slaw.

Related Posts