Washington Lawyer - May/June 2024 - 19

FEATURE
" While we received favorable interest from various federal agencies
and other organizations about the report, the D.C. government never
reached out to us in response, " says Grant Fergusson, EPIC Equal Justice
Works Fellow and a contributor to the report. " To our knowledge, the
D.C. government and its agencies remained silent to the issues addressed
in our report. "
DEVASTATING CONSEQUENCES
In the past decade, technology has evolved by leaps and bounds, with
more far-reaching impacts on everyday life. In the era of artificial intelligence
(AI) and other new technologies, advocates say that basic principles
of human rights should guide the development and use of the resulting
tools.
The Stop Discrimination by Algorithms Act - the first comprehensive
legislation of its type in the country if passed - was introduced
by then-D.C. attorney general Karl Racine in 2021. The act
would make it illegal for businesses and organizations
to use discriminatory algorithms in education,
employment, housing, and public accommodations
and services, including credit,
health care, and insurance.
" Algorithms often rely on personal
traits protected under the D.C. Human
Rights Act. And algorithmic
decision making can amplify discrimination
based on race, gender,
sexual orientation, disability, age,
source of income, credit information,
and other protected traits
when algorithmic models replicate
existing societal inequalities, " the
proposed act states. " Algorithmic decision-making
systems that fail to account
for bias disproportionately harm
marginalized communities. " The bill remains
pending before D.C. Council.
" One of the issues with these algorithms is that oftentimes the incentives
are not aligned to get it right, " De Liban says. " State governments are
happy for benefits to be harder to get and for people to drop off the system.
An algorithm like the one Arkansas used in 2016 allows states to cut
the care of elderly or disabled people to such an extent that they are lying
in their own waste. "
Amos Toh, senior researcher of artificial intelligence and human rights at
Human Rights Watch until March 2024, says the Arkansas cases are illustrative
of how the lack of due diligence in assessing the risks of algorithmic
tools can have serious human rights impacts on the most vulnerable
people.
" That is the very real human consequence of algorithmically enabled cuts
to public services in the U.S., " Toh says. " This algorithm had actually slashed
the home care hours, in very arbitrary ways, that were allocated to people
who were receiving too few hours to begin with. So, you can
imagine some of the devastating consequences for people's
economic and social rights in that context. "
WASHINGTON, D.C.,
WAS USING NEARLY 30 AUTOMATED
DECISION-MAKING SYSTEMS TO SURVEIL,
SCREEN, AND SCORE DISTRICT RESIDENTS
IN AREAS SUCH AS PUBLIC BENEFITS,
HEALTH CARE, POLICING,
AND HOUSING.
In the case of Ethel Jacobs, Legal Aid of Arkansas
filed suit against the state's DHS for violations of the
Due Process Clause of the U.S. Constitution, the Medicaid
Act, and the Americans With Disabilities Act. Legal Aid argued that
the software used by the state was not implemented correctly, mistakenly
putting Jacobs in a category less severe than where she should
have been placed.
" There were multiple errors in the code, such that it was not taking into
account [her] diabetes, which would have given Ms. Jacobs more [inhome
health care] hours, " says Kevin De Liban, founder of TechTonic Justice.
De Liban was director of advocacy at Legal Aid of Arkansas until late
2023.
Jacobs passed away on November 8, 2016, days after the Arkansas
DHS was ordered to restore her in-home health care hours. Legal Aid
continued its fight to protect the rights of Arkansas residents receiving
in-home care, suing the state in 2017 and 2019 for its use of algorithmic
systems.
MECHANISMS FOR MONITORING
Organizations like Human Rights Watch
are closely monitoring the impacts of
AI and emerging technologies. Areas
of particular scrutiny include unemployment
benefits, especially with
dozens of states now using AI to
verify whether applicants are who
they say they are, and labor protections
in the gig economy.
" Unfortunately, this [recognition]
technology does not work very
well, not because the technology
itself is flawed, but because when
the technology is being used in a realworld
context, it is more liable to commit
errors. People, as a result, are shut out
Washington Lawyer - May/June 2024

Table of Contents for the Digital Edition of Washington Lawyer - May/June 2024
