Washington Lawyer - May/June 2024 - 20

FEATURE
Algorithmic Accountability Act of 2023, still pending in subcommittee.
Almost a year earlier, the White House issued the Blueprint for an AI Bill of
Rights anchored on five principles to " guide the design, use, and deployment
of automated systems to protect the American public in the age of
artificial intelligence. " Under the principle of algorithmic discrimination
protections, the framework called upon designers, developers, and deployers
of automated systems to " take proactive and continuous measures
to protect individuals and communities from algorithmic discrimination
and to use and design systems in an equitable way. "
The concern about AI technology is not unique to the United States, Toh
says. " There is widespread reporting, not just by media, but also by the
International Labour Organization, that indicates some very concerning
trends around the prevalence of algorithm-enabled misclassification as
well as some of the techniques to push drivers and workers to work longer
and harder, and sometimes in quite unsafe conditions, " Toh adds.
Among those who raised a flag concerning the potential harms of unregulated
AI was United Nations High Commissioner for Human Rights
Voker Türk, exhorting the Human Rights Council - composed of 47
member states, including the United States - to take a human rightsbased
approach to AI.
" To be effective, to be humane, to put people at the heart of the development
of new technologies, any solution - any regulation - must be
grounded in respect for human rights, " Türk said in July 2023.
IMPACT ON DEMOCRATIC PROCESSES
Another AI-related risk particularly relevant to the United States as it
heads to the polls in November: disinformation and elections.
National Democratic Institute (NDI) senior advisor Daniel Arnaudo says
new AI systems have super-charged the occurrence of " computational
propaganda, " which he defines as the use of both automated and organic
systems - people and bots - to impact political discourse in different
contexts.
When computational propaganda is " essentially changing the field and
changing what the candidates say in different ways, " it impacts voters'
right to participate in an open and democratic process, says Arnaudo.
" The harder thing to get your mind around is also that these systems are
increasingly responsible in different ways for decisions on the back end of
social networks - how the algorithms are sorting your content and presenting
it to you, how they are able to build profiles of individuals that
could then be used to target [people] with political advertising, " he adds.
In addition to studying the integrity of political messaging, NDI supports
domestic observers to better monitor and report on the activity. " We
have really tried to build engagement with tech companies and with the
private sector, " Arnaudo says. " We work on reporting and escalation issues
with Meta and X - even other tech companies that are thinking
about these issues from a cybersecurity perspective, [such as] Microsoft
and Google. We are thinking about how we might train our partners to
keep themselves safe, and that includes informational threats [from generative
AI]. "
NDI has also spearheaded the Design 4 Democracy Coalition, a group of
civic organizations and human rights groups, to ensure that information
20 WASHINGTON LAWYER
* MAY/JUNE 2024
AT THEIR CORE,
THE AI SYSTEMS THAT
GOVERNMENT AGENCIES
USE ARE ASSUMPTION
MACHINES.
technology and social media play a proactive role in supporting democracy
and human rights globally.
" We have been building that network in different ways and using it to
develop consultations on issues, " Arnaudo says. " A lot of it [also] focuses
on how we can get the [tech] companies to abide by better transparency
laws. Or, if not laws, then normative standards. "
HARMFUL ASSUMPTIONS
AI governance is a rising concern globally, so much so that in November
2023 AI technology was a focal point at the United Nations Forum on
Business and Human Rights in Geneva, Switzerland, where panelists
touched upon privacy, implicit bias, the desire for transparency and accountability,
and other issues.
" We are in a rapidly changing environment for this technology. It's being
developed quickly and launched quickly, " Google's human rights and
free expression global policy lead Alexandria Walden said during the forum.
" So, we need more tools as practitioners and companies to help
bring the company along to meet its obligations while we are launching
these products. "
Locally, organizations like EPIC are holding governments to account for
the use and misuse of AI decision-making systems. In September 2023, it
published a follow-up report, " Outsourced & Automated, " that shines a
light on how more and more state and local governments are experimenting
with AI tools created by private companies for greater efficiency
and cost savings.
" At their core, the AI systems that government agencies use are assumption
machines: They make predictions, recommendations, and evaluations
about people based on assumptions gleaned from large databases, "
the report states. " A Medicaid system may assign medical support to
an individual based on what similar individuals typically need. A pretrial
Washington Lawyer - May/June 2024

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