ASET Technology Alberta Fall 2023 - 17

feature
Another concern over effectiveness comes from new developments
in advanced AI. " The outdated aspects refers
primarily to the development of generative AI systems in
the last year or so, " Zajko says, speaking with ASET. Unlike
predictive AI, which uses machine learning algorithms
to look at historical data and make 'predictions' about the
future, generative AI can generate content using that data
and simple prompts.
Responding to ASET about legal responses to training AI
systems to perform unethical activities, a representative
from ISED did say the government is studying the impacts
of this advanced form of the field.
" The increasing prevalence of generative AI has raised a
range of challenges that the government is addressing
through new and existing legislation, " they said.
" The Government of Canada takes the risks posed by AI
systems seriously, and Innovation, Science and Economic
Development Canada officials have been actively engaging
with stakeholders to understand the range of views on
Bill C-27. In fact, the Minister of Innovation, Science and
Industry called a special meeting of the Advisory Council
on Artificial Intelligence on April 14, 2023, to seek advice
of members on the novel challenges posed by generative
AI systems, " They said, adding that the government would
be open to amendments coming from expert testimony that
would improve the Bill.
There are also concerns over AIDA's jurisdiction being limited
to the private sector.
" The AIDA will also ensure accountability for risk associated
with high-impact AI systems used in the course of international
and interprovincial trade and commerce in Canada,
including that of foreign companies, " ISED responded, regarding
foreign actors' use of AI in the country.
" Many of the most 'high impact' applications of AI are those
carried out by governments,
including immigration decisions,
benefits claims, policing and military operations, " Zajko
writes. " Yet, AIDA applies specifically to the private sector.
(A small number of federal government systems have
received algorithmic impact assessments, which have their
own limitations.) "
Regulation of AI usage in the government does exist, under
the purview of another department.
" My concern is that AI cannot and should not be used as
a smokescreen from the real issues of society, " says Santos.
" It's not going to solve all the issues of society. As with
poverty, as with climate change, this by itself does not need
just governance - it needs the interest of the whole public,
and the consciousness that we as society, we have so much
homework to do. "
" Rules for how federal departments and agencies may use
automated systems to inform or make service decisions
may be found in the Treasury Board of Canada Secretariat
(TBS) Directive on Automated Decision Making. TBS has
also developed an Algorithmic Impact Assessment (AIA)
tool to help federal institutions understand and manage
the impacts of their automation projects, "
asked about regulation of AI in government.
ISED said when
" The Government of Canada is committed to implementing
AI in the open and ensuring that the government's use of AI
is governed with clear values, ethics and laws, and in accordance
with human rights, " they said.
These regulations were installed several years ago, though,
and don't apply to many government systems like military.
" For example the RCMP was using Clearview AI, which the
privacy commissioner found to be illegal and mass surveillance, "
Zajko points out, " but they were exempt from the algorithmic
impact assessment. "
While other types of AI in government regulation exist, such
as administrative law principles governing certain procedures
more broadly, Zajko argues these too are not meaty enough.
" There is definitely room for further regulation when it
comes to systems in government, and that's something that
is definitely not addressed by this particular piece of legislation, "
he says.
As with almost any technology, AI is
not inherently problematic. Rather,
it's how we choose to apply it and
regulate it that can bring about
benefits or exacerbate existing
harms. Finally, it's a matter of how
we educate ourselves.
TECHNOLOGY ALBERTA | FALL 2023 | 17

https://www.tbs-sct.canada.ca/pol/doc-eng.aspx?id=32592 https://www.canada.ca/en/government/system/digital-government/digital-government-innovations/responsible-use-ai/algorithmic-impact-assessment.html https://www.priv.gc.ca/en/opc-actions-and-decisions/ar_index/202021/sr_rcmp/#toc1 https://www.priv.gc.ca/en/opc-actions-and-decisions/ar_index/202021/sr_rcmp/#toc1 https://ised-isde.canada.ca/site/advisory-council-artificial-intelligence/en https://ised-isde.canada.ca/site/advisory-council-artificial-intelligence/en

ASET Technology Alberta Fall 2023

Table of Contents for the Digital Edition of ASET Technology Alberta Fall 2023