Well
this month attention has turned from AI being used to create multiple
fake bird species and celebrity images or Microsoft's using
excruciatingly garish alternative landscapes to promote its software -
the focus has shifted back to AI being used by bad actors in global
and domestic political arenas created during election years.
Nature,
WORLD VIEW, 9 April 2024:
Political
candidates are increasingly using AI-generated ‘softfakes’ to
boost their campaigns. This raises deep ethical concerns.
By
Rumman Chowdhury
Of
the nearly two billion people living in countries that are holding
elections this year, some have already cast their ballots. Elections
held in Indonesia and Pakistan in February, among other countries,
offer an early glimpse of what’s in store as artificial
intelligence (AI) technologies steadily intrude into the electoral
arena. The emerging picture is deeply worrying, and the concerns are
much broader than just misinformation or the proliferation of fake
news.
As
the former director of the Machine Learning, Ethics, Transparency and
Accountability (META) team at Twitter (before it became X), I can
attest to the massive ongoing efforts to identify and halt
election-related disinformation enabled by generative AI (GAI). But
uses of AI by politicians and political parties for purposes that are
not overtly malicious also raise deep ethical concerns.
GAI
is ushering in an era of ‘softfakes’. These are images, videos or
audio clips that are doctored to make a political candidate seem more
appealing. Whereas deepfakes (digitally altered visual media) and
cheap fakes (low-quality altered media) are associated with malicious
actors, softfakes are often made by the candidate’s campaign team
itself.
How
to stop AI deepfakes from sinking society — and science
In
Indonesia’s presidential election, for example, winning candidate
Prabowo Subianto relied heavily on GAI, creating and promoting
cartoonish avatars to rebrand himself as gemoy, which means ‘cute
and cuddly’. This AI-powered makeover was part of a broader attempt
to appeal to younger voters and displace allegations linking him to
human-rights abuses during his stint as a high-ranking army officer.
The BBC dubbed him “Indonesia’s ‘cuddly grandpa’ with a
bloody past”. Furthermore, clever use of deepfakes, including an AI
‘get out the vote’ virtual resurrection of Indonesia’s deceased
former president Suharto by a group backing Subianto, is thought by
some to have contributed to his surprising win.
Nighat
Dad, the founder of the research and advocacy organization Digital
Rights Foundation, based in Lahore, Pakistan, documented how
candidates in Bangladesh and Pakistan used GAI in their campaigns,
including AI-written articles penned under the candidate’s name.
South and southeast Asian elections have been flooded with deepfake
videos of candidates speaking in numerous languages, singing
nostalgic songs and more — humanizing them in a way that the
candidates themselves couldn’t do in reality.
What
should be done? Global guidelines might be considered around the
appropriate use of GAI in elections, but what should they be? There
have already been some attempts. The US Federal Communications
Commission, for instance, banned the use of AI-generated voices in
phone calls, known as robocalls. Businesses such as Meta have
launched watermarks — a label or embedded code added to an image or
video — to flag manipulated media.
But
these are blunt and often voluntary measures. Rules need to be put in
place all along the communications pipeline — from the companies
that generate AI content to the social-media platforms that
distribute them.
What
the EU’s tough AI law means for research and ChatGPT
Content-generation
companies should take a closer look at defining how watermarks should
be used. Watermarking can be as obvious as a stamp, or as complex as
embedded metadata to be picked up by content distributors.
Companies
that distribute content should put in place systems and resources to
monitor not just misinformation, but also election-destabilizing
softfakes that are released through official, candidate-endorsed
channels. When candidates don’t adhere to watermarking — none of
these practices are yet mandatory — social-media companies can flag
and provide appropriate alerts to viewers. Media outlets can and
should have clear policies on softfakes. They might, for example,
allow a deepfake in which a victory speech is translated to multiple
languages, but disallow deepfakes of deceased politicians supporting
candidates.
Election
regulatory and government bodies should closely examine the rise of
companies that are engaging in the development of fake media.
Text-to-speech and voice-emulation
software from Eleven Labs, an AI company based in New York City, was
deployed to generate robocalls that tried to dissuade voters from
voting for US President Joe Biden in the New Hampshire primary
elections in January, and to create the softfakes of former
Pakistani prime minister Imran Khan during his 2024 campaign outreach
from a prison cell. Rather than pass softfake regulation on
companies, which could stifle allowable uses such as parody, I
instead suggest establishing election standards on GAI use. There is
a long history of laws that limit when, how and where candidates can
campaign, and what they are allowed to say.
Citizens
have a part to play as well. We all know that you cannot trust what
you read on the Internet. Now, we must develop the reflexes to not
only spot altered media, but also to avoid the emotional urge to
think that candidates’ softfakes are ‘funny’ or ‘cute’. The
intent of these isn’t to lie to you — they are often obviously AI
generated. The goal is to make the candidate likeable.
Softfakes
are already swaying elections in some of the largest democracies in
the world. We would be wise to learn and adapt as the ongoing year of
democracy, with some 70 elections, unfolds over the next few months.
COMPETING
INTERESTS
The
author declares no competing interests.
[my
yellow highlighting]
Charles
Stuart University,
Expert Alert,
media release, 12 April 2024, excerpt:
Governments
must crack down on AI interfering with elections
Charles
Darwin University Computational and Artificial Intelligence expert
Associate Professor Niusha Shafiabady.
Like
it or not, we are affected by what we come across in social media
platforms. The future wars are not planned by missiles or tanks, but
they can easily run on social media platforms by influencing what
people think and do. This applies to election results.
“Microsoft
has said that the election outcomes in India, Taiwan and the US could
be affected by the AI plays by powers like China or North Korea. In
the world of technology, we call this disinformation, meaning
producing misleading information on purpose to change people’s
views. What can we do to fight these types of attacks? Well, I
believe we should question what we see or read. Not everything we
hear is based on the truth. Everyone should be aware of this.
“Governments
should enforce more strict regulations to fight misinformation,
things like: Finding triggers that show signs of unwanted
interference; blocking and stopping the unauthorised or malicious
trends; enforcing regulations on social media platforms to produce
reports to the government to demonstrate and measure the impact and
the flow of the information on the matters that affect the important
issues such as elections and healthcare; and enforcing regulations on
the social media platforms to monitor and stop the fake information
sources or malicious actors.”
The
Conversation,
10 April 2024:
Election
disinformation: how AI-powered bots work and how you can protect
yourself from their influence
AI
Strategist and Professor of Digital Strategy, Loughborough University
Nick Hajli
Social
media platforms have become more than mere tools for communication.
They’ve evolved into bustling arenas where truth and falsehood
collide. Among these platforms, X stands out as a prominent
battleground. It’s a place where disinformation campaigns thrive,
perpetuated by armies of AI-powered bots programmed to sway public
opinion and manipulate narratives.
AI-powered
bots are automated accounts that are designed to mimic human
behaviour. Bots on social media, chat platforms and conversational AI
are integral to modern life. They are needed to make AI applications
run effectively......
How
bots work
Social
influence is now a commodity that can be acquired by purchasing bots.
Companies sell fake followers to artificially boost the popularity of
accounts. These followers are available at remarkably low prices,
with many celebrities among the purchasers.
In
the course of our research, for example, colleagues and I detected a
bot that had posted 100 tweets offering followers for sale.
Using
AI methodologies and a theoretical approach called actor-network
theory, my colleagues and I dissected how malicious social bots
manipulate social media, influencing what people think and how they
act with alarming efficacy. We can tell if fake news was generated by
a human or a bot with an accuracy rate of 79.7%. It is crucial to
comprehend how both humans and AI disseminate disinformation in order
to grasp the ways in which humans leverage AI for spreading
misinformation.
To
take one example, we examined the activity of an account named “True
Trumpers” on Twitter.
The
account was established in August 2017, has no followers and no
profile picture, but had, at the time of the research, posted 4,423
tweets. These included a series of entirely fabricated stories. It’s
worth noting that this bot originated from an eastern European
country.
Research
such as this influenced X to restrict the activities of social bots.
In response to the threat of social media manipulation, X has
implemented temporary reading limits to curb data scraping and
manipulation. Verified accounts have been limited to reading 6,000
posts a day, while unverified accounts can read 600 a day. This is a
new update, so we don’t yet know if it has been effective.
Can
we protect ourselves?
However,
the onus ultimately falls on users to exercise caution and discern
truth from falsehood, particularly during election periods. By
critically evaluating information and checking sources, users can
play a part in protecting the integrity of democratic processes from
the onslaught of bots and disinformation campaigns on X. Every user
is, in fact, a frontline defender of truth and democracy. Vigilance,
critical thinking, and a healthy dose of scepticism are essential
armour.
With
social media, it’s important for users to understand the strategies
employed by malicious accounts.
Malicious
actors often use networks of bots to amplify false narratives,
manipulate trends and swiftly disseminate misinformation. Users
should exercise caution when encountering accounts exhibiting
suspicious behaviour, such as excessive posting or repetitive
messaging.
Disinformation
is also frequently propagated through dedicated fake news websites.
These are designed to imitate credible news sources. Users are
advised to verify the authenticity of news sources by
cross-referencing information with reputable sources and consulting
fact-checking organisations.
Self
awareness is another form of protection, especially from social
engineering tactics. Psychological manipulation is often deployed to
deceive users into believing falsehoods or engaging in certain
actions. Users should maintain vigilance and critically assess the
content they encounter, particularly during periods of heightened
sensitivity such as elections.
By
staying informed, engaging in civil discourse and advocating for
transparency and accountability, we can collectively shape a digital
ecosystem that fosters trust, transparency and informed
decision-making.
Expect
to see AI ‘weaponized to deceive voters’ in this year’s
presidential election
Alfred
Lubrano
As
the presidential campaign slowly progresses, artificial intelligence
continues to accelerate at a breathless pace — capable of creating
an infinite number of fraudulent images that are hard to detect and
easy to believe.
Experts
warn that by November voters in Pennsylvania and other states will
have witnessed counterfeit photos and videos of candidates enacting
one scenario after another, with reality wrecked and the truth nearly
unknowable.
“This
is the first presidential campaign of the AI era,” said Matthew
Stamm, a Drexel University electrical and computer engineering
professor who leads a team that detects false or manipulated
political images. “I believe things are only going to get worse.”
Last
year, Stamm’s group debunked a political ad for then-presidential
candidate Florida Republican Gov. Ron DeSantis ad that appeared on
Twitter. It showed former President Donald Trump embracing and
kissing Anthony Fauci, long a target of the right for his response to
COVID-19.
That
spot was a “watershed moment” in U.S. politics, said Stamm,
director of his school’s Multimedia and Information Security Lab.
“Using AI-created media in a misleading manner had never been seen
before in an ad for a major presidential candidate,” he said.
“This
showed us how there’s so much potential for AI to create voting
misinformation. It could get crazy.”
Election
experts speak with dread of AI’s potential to wreak havoc on the
election: false “evidence” of candidate misconduct; sham videos
of election workers destroying ballots or preventing people from
voting; phony emails that direct voters to go to the wrong polling
locations; ginned-up texts sending bogus instructions to election
officials that create mass confusion.....
Malicious
intent
AI
allows people with malicious intent to work with great speed and
sophistication at low cost, according to the Cybersecurity &
Infrastructure Security Agency, part of the U.S. Department of
Homeland Security.
That
swiftness was on display in June 2018. Doermann’s University of
Buffalo colleague, Siwei Lyu, presented a paper that demonstrated how
AI-generated deepfake videos could be detected because no one was
blinking their eyes; the faces had been transferred from still
photos.
Within
three weeks, AI-equipped fraudsters stopped creating deepfakes based
on photos and began culling from videos in which people blinked
naturally, Doermann said, adding, “Every time we publish a solution
for detecting AI, somebody gets around it quickly.”
Six
years later, with AI that much more developed, “it’s gained
remarkable capacities that improve daily,” said political
communications expert Kathleen Hall Jamieson, director of the
University of Pennsylvania’s Annenberg Public Policy Center.
“Anything we can say now about AI will change in two weeks.
Increasingly, that means deepfakes won’t be easily detected.
“We
should be suspicious of everything we see.”
AI-generated
misinformation helps exacerbate already-entrenched political
polarization throughout America, said Cristina Bicchieri, Penn
professor of philosophy and psychology.
“When
we see something in social media that aligns with our point of view,
even if it’s fake, we tend to want to believe it,” she said.
To
battle fabrications, Stamm of Drexel said, the smart consumer could
delay reposting emotionally charged material from social media until
checking its veracity.
But
that’s a lot to ask.
Human
overreaction to a false report, he acknowledged, “is harder to
resolve than any anti-AI stuff I develop in my lab.
“And
that’s another reason why we’re in uncharted waters.”
