IEEE Technology and Society Magazine - December 2021 - 41

to be accomplished " [28]. A lawyer must " explain a
matter to the extent reasonably necessary to permit
the client to make informed decisions regarding the
representation " [28]. Thus, a lawyer should discuss
with a client the risks and limitations of using an AI
tool for legal-services delivery, if any. Likewise, in
certain circumstances, a lawyer may need to communicate
to the client a decision not to make use
of AI [27].
Some commentators suggest that a lawyer obtain
not only approval, but also " informed consent " from
the client before using AI. But this could create an
unnecessary barrier. Lawyers have great discretion
when choosing the means to accomplish the client's
objectives. If effectiveness testing consistently shows
that an AI tool is reliable and does not present risks
and limitations, arguably neither client approval nor
informed consent should be required. For example,
lawyers surely do not need client approval to
use AI-driven spell checking or electronic research
services. Is client approval required before using AI
for contract review, assuming the cost is the same or
less than human review? If it is shown that contract
review tools are highly effective, client approval is
unlikely to be required. AI tools shown to be effective
for contract review would presumably become
the preferred means for review.
Rule 1.5: Reasonable fees
A lawyer must not charge or collect an unreasonable
fee or an unreasonable amount for expenses. If
a lawyer failed to use technology that would result in
more efficient and effective legal-services delivery, a
client could argue that the lawyer charged an unreasonable
fee in violation of Rule 1.5.
Rule 1.6: Confidentiality
Lawyers must consider their duty to maintain the
confidentiality of client information. If client information
is shared with a third party in connection
with training or developing an AI tool, lawyers must
take appropriate steps to safeguard the client's information
[33]. If a lawyer wishes to use client information
to develop an AI tool, the lawyer should obtain
the client's informed consent and should disclose if
and how the AI tool may be used for other current
and future clients. Complex confidentiality concerns
arise in this situation, and a lawyer must ensure that
the client's confidential information is not disclosed
to current and future clients without first obtaining
December 2021
informed consent. The lawyer must understand
the risks of disclosure, such as the possibility of
reverse-engineering an AI tool to reveal information
provided by the initial client. If the lawyer does not
understand how the AI tool works and its effectiveness,
it will be difficult for the lawyer to adequately
discuss with the client the risks and benefits of using
the AI tool.
Rule 2.1: Exercising independent professional
judgment and providing candid advice
When representing a client, a lawyer must " exercise
independent professional judgment " and " [provide]
candid advice " [30]. The extent of a lawyer's
reliance on AI could raise questions about whether
the lawyer's " professional judgment " remains " independent. "
For example, this issue could arise if a lawyer
incorporates AI outputs into a prediction about
the likelihood of success in a dispute pending in a
particular court. The lawyer's understanding of how
the AI tool works and its effectiveness ought to go
a long way to establishing that the lawyer appropriately
exercised independent professional judgment.
Malpractice liability
Lawyers and other legal-services professionals
could also be subjected to liability for the negligent
use of, or failure to use, AI when delivering legal
services. To recover damages caused by a lawyer's
malpractice, " a plaintiff must demonstrate that the
attorney 'failed to exercise the ordinary reasonable
skill and knowledge commonly possessed by a
member of the legal profession' " [36]. " [A]ttorneys
ordinarily must act consistently with the community
standard of care " [37]. Soft law standards for
effectiveness could contribute to raising community
standards of care regarding lawyers' and other
legal-services professionals' use of AI.
Electronic discovery in law: An
example of AI adoption fostered by
effectiveness testing
When AI tools for the review of electronic documents
were first introduced in the legal industry,
many lawyers expressed skepticism about the
tools themselves and predicted clients and courts
would be slow to accept them. But widespread
adoption of technology-assisted review (TAR)
happened faster than most lawyers expected. This
was due in large part to experts who conducted
41
IEEE Technology and Society Magazine - December 2021

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - December 2021
