Tag: artificial-intelligence

B.C. Wild­fire Ser­vice slams online users for fake images of actual blazes it’s fight­ing

This article was written by Alex Boyd and was published in the Toronto Star on August 10, 2025.

In late July, a truck caught fire near Peach­land, B.C., ignit­ing a blaze that raced up a dry hill­side faster than fire­fight­ers could fol­low, even­tu­ally prompt­ing the clos­ures of two high­ways and evac­u­ations of 400 homes.

Mid­way through the sum­mer fire sea­son, pho­tos of what was dubbed the Drought Hill fire quickly popped up online, show­ing plumes of dull grey smoke rising from the dry scrub­land of the Okanagan Val­ley. But one image, pos­ted on the Face­book page of a self­described “digital cre­ator” the next day, claimed to show the “OUT OF CONTROL” fire in tech­nicolor. The image is so vivid it defies belief — the fluor­es­cent orange flames throw­ing up plumes of coal­black smoke as tiny heli­copters and water bombers whiz over­head.

The image is lush, bright — and fake.

In a rare move, the B.C. Wild­fire Ser­vice spe­cific­ally called out AIgen­er­ated images last week — cit­ing this image and another pur­port­ing to be of a fire near Bear Creek — for spur­ring fear and anxi­ety by spread­ing incor­rect inform­a­tion about the loc­a­tion and beha­viour of the very real fires it is bat­tling. “When mis­in­form­a­tion is spread online, it can start to take root,” says pro­vin­cial inform­a­tion officer Sarah Budd.

“Obvi­ously it’s frus­trat­ing for our staff. We work really hard to quickly and effi­ciently get accur­ate inform­a­tion out.”

Dra­matic pic­tures have a real psy­cho­lo­gical effect on people, says Ali Asgary, a York Uni­versity pro­fessor of dis­aster and emer­gency man­age­ment. The con­cern with overly dram­at­iz­ing or sen­sa­tion­al­iz­ing images is that they might spark panic or unne­ces­sar­ily over­whelm people with inde­cision, he says.

He stresses that gen­er­at­ive AI isn’t inher­ently bad — it’s already show­ing ser­i­ous poten­tial for help­ing those who over­see dis­aster response sift through the flood of data that gushes in dur­ing emer­gency situ­ations, and could even be used for spread­ing edu­ca­tional images or graph­ics that have been veri­fied by pro­fes­sion­als, he says.

But he also warns that there’s a real danger in let­ting just any­one pump out inform­a­tion of dubi­ous ori­gin. For one thing, veri­fy­ing bad inform­a­tion wastes resources that are already stretched thin in an emer­gency situ­ation.

Then there’s the down­stream effect of sow­ing doubt. “Over time, if a lot of this is hap­pen­ing, it cre­ates con­fu­sion for people,” he says. “They don’t really know which sources to fol­low, which sources to trust.

“People start los­ing trust in their agen­cies and people who are sup­posed to provide inform­a­tion to them.”

The image was first pos­ted on an account with the name “Joe­mar Sombero,” a “digital cre­ator” with 43,000 fol­low­ers, on July 31, when the fire had been burn­ing for more than a day. Along­side the dra­matic AI­gen­er­ated photo was an AI­gen­er­ated cap­tion that described the fire as “fast­mov­ing” and “human­caused” and stressed the “massive aer­ial attack under­way” to fight it.

“Sombero” then com­men­ted under­neath: “Stay safe every­one, and huge respect to the fire­fight­ers bat­tling this non­stop!” he wrote, adding a pray­ing hands emoji and the hashtag #Pray­forBC.

Many of the com­ments were furi­ous. “Why are you using AI pic­tures and AI write ups to sen­sa­tion­al­ize Cana­dian Wild­fires? Please stop. It’s not help­ful,” reads one of the first. “Sombero” did not respond to requests for an inter­view the Star sent to the Face­book and asso­ci­ated Ins­tagram accounts, but did hop back into the com­ments to explain their actions.

“The images and write­ups I share are AI­gen­er­ated for illus­trat­ive pur­poses only and are always tagged with a dis­claimer,” they wrote to the com­menter. (When the image was first pos­ted it did not include an AI dis­claimer, accord­ing to the post’s edit his­tory on Face­book; a day later, one was added.) “My goal is never to sen­sa­tion­al­ize but to raise aware­ness about the sever­ity of these events.”

In another com­ment, they say its neces­sary to have more eye­catch­ing images to help people “pay atten­tion in a sea of scrolling.”

In some ways, the account mir­rors the tra­ject­ory of pop­u­lar AI use. It came to life in the spring of 2023 and ini­tially pos­ted mostly nature shots that appear to be reg­u­lar pho­to­graphs, mostly of moun­tains and lakes in the Cana­dian Rock­ies and Cal­gary area.

The poster seems to dis­cover gen­er­at­ive AI around early 2024, and began exper­i­ment­ing with gen­eric Cana­diana con­tent — north­ern lights, moun­tain ranges, people wear­ing plaid — that dis­plays the slightly car­toon­ish look of earlier AI mod­els. A well­received post dur­ing this time period got maybe a few dozen likes or heart emo­jis.

But earlier this year, the account appeared to zero in on nat­ural dis­aster con­tent and begins post­ing increas­ingly real­istic AI­gen­er­ated images of forest fires and vol­ca­noes and in one case, a grizzly bear attack. The AI­gen­er­ated cap­tions got longer, more dra­matic, more emoji­stud­ded.

Fires are a major focus. There’s a AI image of the fires in La Loche, Sask., and fake images of people who evac­u­ated ahead of blazes in Cali­for­nia. (“The after­math … is heart­break­ing,” the cap­tion notes.)

But its earth­quakes that seem to be par­tic­u­larly enti­cing to social media eye­balls.

Last week, an AI­image and post about how “THE EARTH’S CRUST WON’T STAY QUIET” got more than 1,000 inter­ac­tions, as did one about “SHAKING” in South­ern Cali­for­nia. Then a big win. A single post about the num­ber of earth­quakes world­wide in the last month — “EARTH IS SHAKING HARD” — saw more than 6,000 people hit the react but­ton.

A short video pos­ted to the account fea­tures a cas­cade of hap­pylook­ing emo­jis with a cel­eb­rat­ory title: “I got over 6,500 reac­tions on one of my posts last week! Thanks every­one for your sup­port!”

Another term for the flow of fake imagery being churned out by the account here is AI slop, says Lauren Dwyer, an asso­ciate pro­fessor at Mount Royal Uni­versity who stud­ies emer­ging tech­no­lo­gies and how they influ­ence beha­viour.

“It’s just AI­gen­er­ated images for the sake of AI­gen­er­ated images,” she says.

Even the poster’s responses to com­menters have the fawn­ing tone of Chat­GPT, she points out. (“I hear your frus­tra­tion,” he assures one per­son upset about his AI use.) It’s pos­sible this is one per­son let­ting gen­er­at­ive AI run amok, she says, but it’s also pos­sible the account is run by bots. Either way, these accounts are often set up less to spread reli­able inform­a­tion than to gen­er­ate clicks.

Much like how snack foods are engin­eered to keep you eat­ing, AI is also trained to ensnare your atten­tion. “AI cre­ates really, really eye­c­atch­ing images,” Dwyer says.

“Get­ting that bal­ance of light­ing and drama, all of the aspects that go into incred­ible photo journ­al­ism? To be able to do that with a click, as opposed to hav­ing to find the right angle and nav­ig­ate wild­fire smoke, it just removes all the bar­ri­ers.”

The gold rush to get social media engage­ment is why AI­images of everything from nat­ural dis­asters to celebrit­ies behav­ing badly to weirdly unap­peal­ing recipes have flooded social media.

Face­book itself allows accounts with a large num­ber of fol­low­ers to make money through ads in videos and paid sub­scribers. It’s not clear if this is the page’s goal — it cur­rently appears to have two sub­scribers pay­ing $1.29 a month, one of whom is a Toronto Star reporter — but it provides a win­dow into the incent­ives at play.

“It’s a busi­ness,” Dwyer says. “Every­one who is doing it is tak­ing up space in a media land­scape and try­ing to find their niche.

“Like, it’s about try­ing to make money, and it doesn’t really mat­ter how that gets done.”