HLA Magazine Feature: Personal Injury Update – MARCH 2024
Its spring – and a young lawyer’s heart of course, turns
to case law. I, however, have two cases
that will break that young lawyer’s heart.
The first case concerns the perennial problem of limitations. The second case relates to a lawyer in B.C. who
touched the hot stove of artificial intelligence and got burned.
Sanei v. Debarros, 2024 ONCA 104
On February 9th the Court of Appeal for
Ontario released a judgment, dismissing a plaintiff’s appeal of an order dismissing
his action on summary judgment for missing a limitation period. The plaintiff / appellant was injured in an
auto accident on February 2, 2013, but did not issue his action until March 3,
2016. The motions judge, based upon the
medical evidence and the evidence of the plaintiff, concluded that the
plaintiff had suffered “serious and permanent injuries” in the MVA which were
“discoverable within the two-year limitation period”. As a result, the motions
judge concluded that there was no genuine issue for trial and consequently
dismissed the action.
In his appeal the plaintiff / appellant asserted that
the motions judge erred in three ways:
- That the motions judge in
determining that the serious and permanent nature of the plaintiff’s injuries
was discoverable within the two-year period failed to specify on what date
within that two-year period the injuries were discoverable. - That there was no evidence demonstrating that
the appellant discovered that he had a serious and permanent impairment prior
to March 2, 2014 (two years before the date the claim was issued). - That the motion judge applied the wrong
threshold test, characterizing the appellant’s injuries as “serious and
permanent injuries” rather than considering whether they constituted “serious
and permanent impairment” as defined under O. Reg. 461/96 which defines the
proof that must be adduced at a “threshold motion” showing that the plaintiff
meets the requirements of s. 267.5 of the Insurance Act.
After a
careful review of the operation of section 5 of the Limitations Act, the
Court of Appeal rejected each ground of appeal advanced. The Court of Appeal took note of the issue of
determining when a person knows or ought reasonably to know that they have a
potential cause of action premised upon the suffering of an injury, that
section 5(2) has the effect of creating a presumption that the plaintiff knew
of their loss on the day the impugned act or omission they allege caused the
loss “unless the contrary is proved”.
The court then canvassed the leading cases on point.[1]
In
addressing the first of three grounds advanced the Court of Appeal acknowledged
that the motions judge did not identify a specific date upon which the
plaintiff was aware of the severity and permanence of his impairment. However, that was not a reversable error
because he referenced the correct analysis under Piexiero and tethered
himself to the available evidence which was demonstrative of the fact that the
plaintiff either knew or ought to have known that “a claim would be an
appropriate remedy for the loss” prior to the commencement of 2 years leading
up to the issuance of the action.[2]
On
the second ground advanced, the Court of Appeal in review of the Motions
Judge’s reasons noted that he clearly canvassed medical evidence proffered
showing that the plaintiff had disclosed, prior to the two-year period
(commencing March 2, 2014) leading up to the issuance of the Claim (March 2,
2016) significant health concerns stemming from the accident:
“In
particular, the motion judge noted the December 19, 2013, report of
psychiatrist Dr. Azadian, which concluded that the appellant suffered from a
serious psychological disorder…”.[3]
The
Court noted that the December 19, 2013, report contained claims of injuries
which continued and were echoed in subsequent medical reports and remained
largely unchanged up to and past the date of the statement of claim being
issued.
With
respect to the third ground, the Court of Appeal indicated that: “it is not the
policy of the law or the intent of the limitations provisions to require people
to commence actions before they know that they have a substantial chance to
succeed in recovering a judgment for damages.”[4]
Moreover, the Court advised that
establishing such a high threshold of knowledge on the plaintiff’s part for the
discoverability clock to begin to run is
too extreme a position for the Court to accept and would essentially “move the needle too close to certainty”,
which is not the standard required to trigger discovery of a claim and the
commencement of the limitation period.[5]
Zhang v. Chen, 2024 BCSC
285
This
is not a personal injury case but rather a family law case. However, it is an important case for everyone
to know with respect to the use, misuse and negligent use of artificial
intelligence (AI) when doing legal research.
It is a warning to not proffer a case found using AI as judicial
authority for a proposition without verifying its authenticity yourself.
In
this case, a father and mother had a high conflict separation. They disputed their parenting time and other
matters of access to the children. At a
contentious juncture, the father felt he was being unreasonably denied access
to his children so he instructed his counsel to bring an application to address
his grievance. Counsel for the father,
in his materials submitted and relied upon two cases – and only two cases in
support of his application: “M.M. v. A.M., 2019 BCSC 2060” and “B.S.
v. S.S., 2017 BCSC 2162”.
Counsel
for the mother in preparing a response to the application contacted counsel for
the father seeking copies of those cases as they could not locate and verify
the cases. After some back and forth,
counsel for the father advised the court and counsel in an email:
I made a
serious mistake when preparing a recent Notice of Application for my client,
Mr. Wei Chen, by referring to two cases suggested by Chat GTP (an artificial
intelligent tool) without verifying the source of information. I had no idea
that these two cases could be erroneous. After my colleague pointed out the
fact that these could not be located, I did research of my own and could not
detect the issues either. Regardless of the level of reliability of Al aids, I
should have used more reliable platforms for doing legal research and should
have verified the source of information that was going to be presented in court
and/or exchanged with the opposing counsel. I have taken this opportunity to
review the relevant professional codes of conduct and reflected on my action. I
will not repeat the same mistake again. I had no intention to mislead the
opposing counsel or the court and sincerely apologize for the mistake that I
made.
….
I alone made
the mistake with respect to the erroneous citations and nobody else in my
office was any part of it.[6]
At
the hearing of the matter, after receiving this acknowledgement, the court
ruled on the substance of the matter before it.
It then gave notice to counsel to prepare submissions with respect to
the issue of costs of the matter being borne personally by the lawyer.
After
hearing submissions, the Court stated:
Citing fake
cases in court filings and other materials handed up to the court is an abuse
of process and is tantamount to making a false statement to the court.
Unchecked, it can lead to a miscarriage of justice.[7]
The
Court further went on to state:
The risks of
using ChatGPT and other similar tools for legal purposes was recently
quantified in a January 2024 study: Matthew Dahl et. al., “Large Legal
Fictions: Profiling Legal Hallucinations in Large Language Models” (2024)
arxIV:2401.01301. The study found that legal hallucinations are alarmingly
prevalent, occurring between 69% of the time with ChatGPT 3.5 and 88% with
Llama 2. It further found that large language models (“LLMs”) often fail to
correct a user’s incorrect legal assumptions in a contrafactual question setup,
and that LLMs cannot always predict, or do not always know, when they are
producing legal hallucinations. The study states that “[t]aken together, these
findings caution against the rapid and unsupervised integration of popular LLMs
into legal tasks.”[8]
Ultimately,
the Court accepted that counsel’s error was not intentional but
unintentional. Regardless, her error
precipitated significant excess costs in the proceeding which her client should
not have been responsible for. She was
ordered to do two things. One, pay costs
personally and two, review all files she has before the Court to see if she has
used AI to submit hallucinated cases in other matters and advise opposing
counsel and the Court if she has.
In a final
comment, the Court made a statement we should all remember:
As this case
has unfortunately made clear, generative AI is still no substitute for the
professional expertise that the justice system requires of lawyers. Competence
in the selection and use of any technology tools, including those powered by
AI, is critical. The integrity of the justice system requires no less.[9]
[1] Morrison v. Barzo, 2018
ONCA 979, 144 O.R. (3d) 600 Peixeiro v. Haberman, [1997] 3 S.C.R. 549; Dubreuil
v. Lalande, 2014 ONSC 7433; Rockford v. Haque, 2019 ONSC 474; and Yasmin
v. Alexander, 2016 ONCA 165.
[2] Sanei, at para. 17
[3] Ibid, at para. 20
[4] Ibid, para. 24
[5] For this proposition, the Court of Appeal cited: Grant Thornton
LLP v. New Brunswick, 2021 SCC 31, at paras. 47-48, citing K.L.B. v.
British Columbia, 2003 SCC 51, at para. 55.
[6] Zhang v. Chen, 2024 BCSC 285, para. 12.
[7] Ibid, para. 29
[8] Ibid, para. 38
[9] Ibid, para. 46.