August 12, 2025 • 12 mins
Aug. 12, 2025 - Artificial intelligence is invading New York's political campaigns this year and the burgeoning technology could be even more prevalent in the future, according to Annie McDonough, a senior reporter for City & State New York. This is an excerpt of Dispatches from Planet Albany.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:04):
This is WCNYS the Capitol press Room, and we're going
to play an excerpt of our dispatches from Planet Albany.
Conversation with Annie McDonough, a senior city hall reporter for
City in State New York who joined us this week
to talk about the use of artificial intelligence by political campaigns.
For starters, can you talk a little bit about how
AI might be used behind the scenes in a way
(00:26):
that campaign followers voters might never know unless there's reporting
on it, or maybe a campaign is forced to disclose it.
Speaker 2 (00:36):
Yeah.
Speaker 3 (00:36):
Yeah, I kind of put it in more in two camps.
There's the sort of outfront self disclosure public uses of
AI that either you're told about because the campaign says
this is AI produced or because it looks and just.
Speaker 2 (00:48):
Kind of like smacks of AI content.
Speaker 3 (00:51):
The behind the scenes uses are a bit more interesting
to me because when you can't really tell for sure
if what you're looking at or if the you know,
campaign or fundraising email that you're looking at is AI
generated or generated help of AI, then it kind of
leads to more questions.
Speaker 2 (01:07):
About what is being done by the campaign.
Speaker 3 (01:11):
So it might be fundraising emails, it might be regular
sort of communications statements that are being sent out by
a campaign.
Speaker 2 (01:18):
Or a candidate. That whole world of.
Speaker 3 (01:20):
The behind the scenes usage is still developing, and we
don't have like a firm grasp on everywhere that's being
used so far. But I'm curious to see, you know,
in New York, if candidates are using it to help
organize canvassers or volunteers. In other states, other cities, we've
seen it used for like robocalls for volunteers, canvassing, like
(01:44):
fundraising calls, things like that.
Speaker 2 (01:46):
In New York, I don't think, or at least that
we're aware of.
Speaker 3 (01:50):
We haven't seen it used in that way super broadly yet,
but we've seen also falling in the behind the scenes
category that does later come to light are things like
deep fakes.
Speaker 2 (02:01):
A couple of years ago, there was that Keith Wright.
Speaker 3 (02:05):
Yes, purported to be an audio recording of him criticizing
arrival in Harlem, but was in fact a.
Speaker 2 (02:11):
Deep fake of his voice.
Speaker 3 (02:13):
I think deep fakes are in the category of the
behind the scenes discrete uses that you absolutely want to
have come to light because they are very manipulative and
trying to fool voters.
Speaker 2 (02:25):
Or journalists or whoever.
Speaker 3 (02:26):
It is, but they are also a little bit less
nefarious behind these scenes uses that's just like done for
economy or efficiency. You know, a scrappy, low staffed campaign
is using it for fundraising copy that kind of thing.
So it kind of runs the gamut. But I think
the whole spectrum of it we're all still trying to
(02:47):
get a handle on.
Speaker 1 (02:48):
So I was out myself as an artificial intelligence novice.
But when you talk about external communications, like a fundraising email,
does that just mean dumping all this different information into
chat GPT and seeing what comes out.
Speaker 3 (03:01):
Yeah, I mean that's what I'm told is that you
can kind of put we want to raise We've got
this fundraising deadline, we want to raise one hundred k
I don't know from donors, put in all the requisite
details about CFB deadlines, and split something out that sounds
like it's coming from a political candidate. Maybe it's the job,
but like an interim would have done in the past,
(03:22):
or a low level staffer, and they're able to just
kind of generate it.
Speaker 2 (03:26):
Faster and a bit. I guess maybe like snappy er
is the idea.
Speaker 3 (03:35):
I haven't seen direct examples of this, but I've been
told by consultants that that's a pretty like low risk
usage of generative AI like chat GPT, And then I
should mention we've also seen it like pretty notably this
past cycle in the city election, in helping to research
(03:55):
policy that was a big Cuomo campaign, Like was that
in the footnote of his housing policy he released back
in the spring, there was a mention of chat gpt
in a link in a footnote that they would later
cop to having used chatgbt to help research housing policies
(04:17):
or housing issues I guess around the city to help
craft this policy. Obviously, what they said was that it
was a research tool, it wasn't written by AI, and
that's how they explained that. But I think that shows
that campaigns, staffers, researchers, even like higher level consultants are
(04:38):
maybe helping synthesize a Google search with chatgybt to kind
of deliver the bullet points of one of the biggest
housing issues of the city. What are the stats on
like the creation of units in the last ten years,
that kind of thing that I guess people are just
getting done faster with AI. And I'll admit to not
being a user of chatgbt myself very.
Speaker 2 (04:59):
Often so I don't.
Speaker 3 (05:01):
Know how much faster it actually works, but I think
that's the idea.
Speaker 1 (05:05):
Do you use the word debacle to describe the Cuomo
campaign experience this year? And while yes, it made it
into some political newsletters, I think people following the campaign
were aware of it. People who hate Andrew Cuomo probably
cited it as some reason not to vote for him.
But do you think people in general? Do you think
(05:27):
voters in general care about this example of AI being
used or AI being deployed more broadly in campaigns.
Speaker 3 (05:36):
I think I'll do this example at first. I think
it is a more sort of clear division of the
like consultant class, and media, and like that whole division
of people that are following this very closely in Queen
who are not fans of Andrew Cuomo, who care a
lot more about this footnote in the housing policy than
(05:56):
what the housing policy would actually do. I think in general,
role no, like the average New Yorker probably doesn't care
too much about this or even heard about it. What
it did do for Cuomo, though, was takable out of
the oxygen of like a.
Speaker 2 (06:09):
Couple of days of a news cycle.
Speaker 3 (06:11):
And put a lot more of the attention on this
like really proofreading error probably from their view instead of
what the policy actually included.
Speaker 1 (06:22):
Yeah, it becomes a process story. It becomes a critique
of his campaign more broadly. That's a good point.
Speaker 2 (06:29):
Yeah.
Speaker 3 (06:29):
And also like that was in a string of other
sort of campaign mishaps and like dumb errors that had
been made.
Speaker 1 (06:35):
Yeah, campaign finance filings, they're so hard.
Speaker 3 (06:38):
But I think what I'm interested in is the extent
to which just like the average voter or observer of
politics and somebody's like actually deciding who to vote for,
is turned off or not by obviously AI generated content.
(06:59):
I think they're probably not gonna be thinking that much
if they even are told about it about the behind
the scenes use as a AI, unless it's something goes
really wrong or something is disclosed by accident by a
campaign as to have been AI generated.
Speaker 2 (07:11):
But the like the.
Speaker 3 (07:13):
Cartoon generated images, the like sort of uncanny Valley like
music videos that candidates put out, the stuff that you
see Donald Trump do a lot of the like Maga
crowd very readily embraces the AI generated stuff. Does it
strike people as strange and an authentic to to strike
them as funny and something that they're seeing more of
(07:35):
in their everyday life and are happy to see sometimes
like out of touch politicians embrace Now.
Speaker 2 (07:41):
I think it depends on the voter.
Speaker 1 (07:43):
I think it depends on who it's attacking. If you're
a Trump voter, Trump's making it, it's great. If you're
a Democrat and there's a video like there was attacking
a least dephonic, a North Country Republican, and you're like, well,
this is hilarious, this is on the money.
Speaker 3 (07:55):
Yeah, yeah, yeah, yeah, Maybe just the same audience as
gravitate to whoever or is putting out that content, but.
Speaker 2 (08:02):
They would anyways.
Speaker 3 (08:04):
But yeah, there is like an element of trustworthiness maybe
and authenticity that I do think genuinely like on the
fence voters might look for in their candidates. And so
I wonder if people are turned off by that. It's
all still very new, and I don't think we have
a good sense of like how people generally feel about it.
But you're probably right that people gravitate to their candidates
(08:26):
and whatever they're doing, whatever that is.
Speaker 1 (08:28):
Either way, although the way you describe it is an
issue that maybe gets someone off the fence, so to speak,
is something that speaks to me is being more of
an issue maybe in a primary where maybe I kind
of like all these different people, but one person kind
of give me the ick because of the way they
used AI. So maybe in that way that intra or
(08:52):
inter I can never remember the correct use of party violence.
Maybe they don't want to see AI. But like you said,
I think it's definitely one of those issues where we
have to see how it plays out. But as it
plays out, what rules are there, if any, for campaigns
to follow. Do they need to be explicit about the
use of AI for video? Do they need to include
(09:14):
a footnote in an email if chat GPT is used?
And are there differences depending on what you're running for,
whether it's like a city election versus state versus congress.
Speaker 2 (09:26):
There's not the short answers that there's not a whole
lot of regulation. Oh good, I think, yeah. You always
love to hear that.
Speaker 3 (09:32):
The law that's on the books that's probably the closest
to having to do with the specific use of AI
and political communications and by campaigns is a state requirement
I think is.
Speaker 2 (09:42):
A year old or so.
Speaker 3 (09:44):
It's a disclosure requirement for what's called materially deceptive media
in political communications. Basically sort of overarching covering AI generated
content and deep fakes in addition to having to disclose
that information in a video or audio or text ad.
(10:05):
Candidates who have their likeness used have a right to
action through that requirement if it's not disclosed. So there
is some language around that in in state law right
now to cover candidates across the state. But I don't
think we've really seen that put to use or put
to the test in a major way that I'm aware of,
(10:26):
at least, And so I think that is one sort
of like to be determined how that's going to work
aspect of it. The other is that that deep fake
sort of focused regulation is just one aspect of how
AI is used in campaigns, So it doesn't cover the
things like if AI is used to research a policy
(10:49):
or write a statement. It's more about how a campaign's
like targets arrival or another candidate. So a lot of
other uses of it are so kind of open and
up for grabs. A couple of states have more explicit
prohibitions on deep fakes. There's other legislation that's pending in
(11:10):
the state for sort of non campaign related uses of AI.
And I know you guys are following like the state
level stuff on the RAYS Act, there's city legislation that
would I think there's a city bill that would have
campaigns or have candidate allow candidates to go to an
(11:31):
AI company instead and say that they don't allow their
likeness to be used in another person's ad, and then
create a right traction for that too, But that hasn't
been there's been much.
Speaker 2 (11:42):
Action around that either.
Speaker 3 (11:44):
I think with all AI regulation, it's all very new,
and there's a pretty active lobby gearing up to spend
against anything that is going to be sort of a
meaningful restriction on how these companies operate. So it's really
in its infancy any kind of meaningful restriction on how
campaigns u us A stuff.
Speaker 1 (12:04):
And that was Annie McDonough, a senior City Hall reporter
for City and State New York. And you can hear
our full conversation on dispatches from Planet Albany at Capitolpressroom
dot org or wherever you download your favorite podcasts. If
you're enjoying the Capitol Pressroom, please consider making a recurring
contribution to help make our journalism possible. To show your support,
(12:26):
visit Capitol Pressroom dot org. Slash donate,
This is WCNYS the Capitol press Room, and we're going
to play an excerpt of our dispatches from Planet Albany.
Conversation with Annie McDonough, a senior city hall reporter for
City in State New York who joined us this week
to talk about the use of artificial intelligence by political campaigns.
For starters, can you talk a little bit about how
AI might be used behind the scenes in a way
(00:26):
that campaign followers voters might never know unless there's reporting
on it, or maybe a campaign is forced to disclose it.
Speaker 2 (00:36):
Yeah.
Speaker 3 (00:36):
Yeah, I kind of put it in more in two camps.
There's the sort of outfront self disclosure public uses of
AI that either you're told about because the campaign says
this is AI produced or because it looks and just.
Speaker 2 (00:48):
Kind of like smacks of AI content.
Speaker 3 (00:51):
The behind the scenes uses are a bit more interesting
to me because when you can't really tell for sure
if what you're looking at or if the you know,
campaign or fundraising email that you're looking at is AI
generated or generated help of AI, then it kind of
leads to more questions.
Speaker 2 (01:07):
About what is being done by the campaign.
Speaker 3 (01:11):
So it might be fundraising emails, it might be regular
sort of communications statements that are being sent out by
a campaign.
Speaker 2 (01:18):
Or a candidate. That whole world of.
Speaker 3 (01:20):
The behind the scenes usage is still developing, and we
don't have like a firm grasp on everywhere that's being
used so far. But I'm curious to see, you know,
in New York, if candidates are using it to help
organize canvassers or volunteers. In other states, other cities, we've
seen it used for like robocalls for volunteers, canvassing, like
(01:44):
fundraising calls, things like that.
Speaker 2 (01:46):
In New York, I don't think, or at least that
we're aware of.
Speaker 3 (01:50):
We haven't seen it used in that way super broadly yet,
but we've seen also falling in the behind the scenes
category that does later come to light are things like
deep fakes.
Speaker 2 (02:01):
A couple of years ago, there was that Keith Wright.
Speaker 3 (02:05):
Yes, purported to be an audio recording of him criticizing
arrival in Harlem, but was in fact a.
Speaker 2 (02:11):
Deep fake of his voice.
Speaker 3 (02:13):
I think deep fakes are in the category of the
behind the scenes discrete uses that you absolutely want to
have come to light because they are very manipulative and
trying to fool voters.
Speaker 2 (02:25):
Or journalists or whoever.
Speaker 3 (02:26):
It is, but they are also a little bit less
nefarious behind these scenes uses that's just like done for
economy or efficiency. You know, a scrappy, low staffed campaign
is using it for fundraising copy that kind of thing.
So it kind of runs the gamut. But I think
the whole spectrum of it we're all still trying to
(02:47):
get a handle on.
Speaker 1 (02:48):
So I was out myself as an artificial intelligence novice.
But when you talk about external communications, like a fundraising email,
does that just mean dumping all this different information into
chat GPT and seeing what comes out.
Speaker 3 (03:01):
Yeah, I mean that's what I'm told is that you
can kind of put we want to raise We've got
this fundraising deadline, we want to raise one hundred k
I don't know from donors, put in all the requisite
details about CFB deadlines, and split something out that sounds
like it's coming from a political candidate. Maybe it's the job,
but like an interim would have done in the past,
(03:22):
or a low level staffer, and they're able to just
kind of generate it.
Speaker 2 (03:26):
Faster and a bit. I guess maybe like snappy er
is the idea.
Speaker 3 (03:35):
I haven't seen direct examples of this, but I've been
told by consultants that that's a pretty like low risk
usage of generative AI like chat GPT, And then I
should mention we've also seen it like pretty notably this
past cycle in the city election, in helping to research
(03:55):
policy that was a big Cuomo campaign, Like was that
in the footnote of his housing policy he released back
in the spring, there was a mention of chat gpt
in a link in a footnote that they would later
cop to having used chatgbt to help research housing policies
(04:17):
or housing issues I guess around the city to help
craft this policy. Obviously, what they said was that it
was a research tool, it wasn't written by AI, and
that's how they explained that. But I think that shows
that campaigns, staffers, researchers, even like higher level consultants are
(04:38):
maybe helping synthesize a Google search with chatgybt to kind
of deliver the bullet points of one of the biggest
housing issues of the city. What are the stats on
like the creation of units in the last ten years,
that kind of thing that I guess people are just
getting done faster with AI. And I'll admit to not
being a user of chatgbt myself very.
Speaker 2 (04:59):
Often so I don't.
Speaker 3 (05:01):
Know how much faster it actually works, but I think
that's the idea.
Speaker 1 (05:05):
Do you use the word debacle to describe the Cuomo
campaign experience this year? And while yes, it made it
into some political newsletters, I think people following the campaign
were aware of it. People who hate Andrew Cuomo probably
cited it as some reason not to vote for him.
But do you think people in general? Do you think
(05:27):
voters in general care about this example of AI being
used or AI being deployed more broadly in campaigns.
Speaker 3 (05:36):
I think I'll do this example at first. I think
it is a more sort of clear division of the
like consultant class, and media, and like that whole division
of people that are following this very closely in Queen
who are not fans of Andrew Cuomo, who care a
lot more about this footnote in the housing policy than
(05:56):
what the housing policy would actually do. I think in general,
role no, like the average New Yorker probably doesn't care
too much about this or even heard about it. What
it did do for Cuomo, though, was takable out of
the oxygen of like a.
Speaker 2 (06:09):
Couple of days of a news cycle.
Speaker 3 (06:11):
And put a lot more of the attention on this
like really proofreading error probably from their view instead of
what the policy actually included.
Speaker 1 (06:22):
Yeah, it becomes a process story. It becomes a critique
of his campaign more broadly. That's a good point.
Speaker 2 (06:29):
Yeah.
Speaker 3 (06:29):
And also like that was in a string of other
sort of campaign mishaps and like dumb errors that had
been made.
Speaker 1 (06:35):
Yeah, campaign finance filings, they're so hard.
Speaker 3 (06:38):
But I think what I'm interested in is the extent
to which just like the average voter or observer of
politics and somebody's like actually deciding who to vote for,
is turned off or not by obviously AI generated content.
(06:59):
I think they're probably not gonna be thinking that much
if they even are told about it about the behind
the scenes use as a AI, unless it's something goes
really wrong or something is disclosed by accident by a
campaign as to have been AI generated.
Speaker 2 (07:11):
But the like the.
Speaker 3 (07:13):
Cartoon generated images, the like sort of uncanny Valley like
music videos that candidates put out, the stuff that you
see Donald Trump do a lot of the like Maga
crowd very readily embraces the AI generated stuff. Does it
strike people as strange and an authentic to to strike
them as funny and something that they're seeing more of
(07:35):
in their everyday life and are happy to see sometimes
like out of touch politicians embrace Now.
Speaker 2 (07:41):
I think it depends on the voter.
Speaker 1 (07:43):
I think it depends on who it's attacking. If you're
a Trump voter, Trump's making it, it's great. If you're
a Democrat and there's a video like there was attacking
a least dephonic, a North Country Republican, and you're like, well,
this is hilarious, this is on the money.
Speaker 3 (07:55):
Yeah, yeah, yeah, yeah, Maybe just the same audience as
gravitate to whoever or is putting out that content, but.
Speaker 2 (08:02):
They would anyways.
Speaker 3 (08:04):
But yeah, there is like an element of trustworthiness maybe
and authenticity that I do think genuinely like on the
fence voters might look for in their candidates. And so
I wonder if people are turned off by that. It's
all still very new, and I don't think we have
a good sense of like how people generally feel about it.
But you're probably right that people gravitate to their candidates
(08:26):
and whatever they're doing, whatever that is.
Speaker 1 (08:28):
Either way, although the way you describe it is an
issue that maybe gets someone off the fence, so to speak,
is something that speaks to me is being more of
an issue maybe in a primary where maybe I kind
of like all these different people, but one person kind
of give me the ick because of the way they
used AI. So maybe in that way that intra or
(08:52):
inter I can never remember the correct use of party violence.
Maybe they don't want to see AI. But like you said,
I think it's definitely one of those issues where we
have to see how it plays out. But as it
plays out, what rules are there, if any, for campaigns
to follow. Do they need to be explicit about the
use of AI for video? Do they need to include
(09:14):
a footnote in an email if chat GPT is used?
And are there differences depending on what you're running for,
whether it's like a city election versus state versus congress.
Speaker 2 (09:26):
There's not the short answers that there's not a whole
lot of regulation. Oh good, I think, yeah. You always
love to hear that.
Speaker 3 (09:32):
The law that's on the books that's probably the closest
to having to do with the specific use of AI
and political communications and by campaigns is a state requirement
I think is.
Speaker 2 (09:42):
A year old or so.
Speaker 3 (09:44):
It's a disclosure requirement for what's called materially deceptive media
in political communications. Basically sort of overarching covering AI generated
content and deep fakes in addition to having to disclose
that information in a video or audio or text ad.
(10:05):
Candidates who have their likeness used have a right to
action through that requirement if it's not disclosed. So there
is some language around that in in state law right
now to cover candidates across the state. But I don't
think we've really seen that put to use or put
to the test in a major way that I'm aware of,
(10:26):
at least, And so I think that is one sort
of like to be determined how that's going to work
aspect of it. The other is that that deep fake
sort of focused regulation is just one aspect of how
AI is used in campaigns, So it doesn't cover the
things like if AI is used to research a policy
(10:49):
or write a statement. It's more about how a campaign's
like targets arrival or another candidate. So a lot of
other uses of it are so kind of open and
up for grabs. A couple of states have more explicit
prohibitions on deep fakes. There's other legislation that's pending in
(11:10):
the state for sort of non campaign related uses of AI.
And I know you guys are following like the state
level stuff on the RAYS Act, there's city legislation that
would I think there's a city bill that would have
campaigns or have candidate allow candidates to go to an
(11:31):
AI company instead and say that they don't allow their
likeness to be used in another person's ad, and then
create a right traction for that too, But that hasn't
been there's been much.
Speaker 2 (11:42):
Action around that either.
Speaker 3 (11:44):
I think with all AI regulation, it's all very new,
and there's a pretty active lobby gearing up to spend
against anything that is going to be sort of a
meaningful restriction on how these companies operate. So it's really
in its infancy any kind of meaningful restriction on how
campaigns u us A stuff.
Speaker 1 (12:04):
And that was Annie McDonough, a senior City Hall reporter
for City and State New York. And you can hear
our full conversation on dispatches from Planet Albany at Capitolpressroom
dot org or wherever you download your favorite podcasts. If
you're enjoying the Capitol Pressroom, please consider making a recurring
contribution to help make our journalism possible. To show your support,
(12:26):
visit Capitol Pressroom dot org. Slash donate,
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.