The coronavirus has created a secondary scourge of online scamming that is testing all the biggest companies on the internet, from Facebook and Google to Amazon, Twitter and even Snapchat.

Ads for products like face masks that make aggressive claims about staving off infection are all over the internet. They are preying on consumers through apps, search and e-commerce channels, despite the best high-tech efforts of platforms to snuff out the baddest actors in a game of digital whack-a-mole.

This week, an investigation into one such company served as yet another reminder that, armed with the most basic internet advertising tools, opportunists are capitalizing on the pandemic. The company, registered under the name Novads OU, with a mailing address in Estonia, has been hawking a product called OxyBreath Pro, which makes claims of blocking airborne viruses. It’s the type of product that is familiar to almost everyone stuck at home browsing the internet during the lockdown, encountering ads for other similar snakeoil products that tout COVID treatments, prevention and other dubious promotions.

Online security experts say that the claims around OxyBreath Pro masks smack of false advertising. “This appears to be a network of e-commerce sites, companies and affiliate marketers peddling fake, defective or perhaps even nonexistent products to defraud unwary users,” says Jordan Herman, threat researcher at RiskIQ, which helped Ad Age analyze the company’s online practices. 

The company has been raising red flags overseas since at least March, when the U.K.’s Advertising Standards Authority said Novads OU was among the companies running “misleading” ads related to the pandemic, according to a BBC report.

However, even with the official alarm, ads for OxyBreath Pro have popped up on major websites, including Google, Bing, Facebook, Instagram and Snapchat. Novads OU also has its digital fingerprints on a string of websites with domain names like “hyperstech,” “hypertechz,” “drone-x,” “daily gadgets advisor,” and appeared to run Facebook pages that promoted its signature product.

An attempt to reach Novads OU through the email address on the websites was unsuccessful.

Some of the websites lead to links to buy the product on Amazon, where the reviews for the masks were mostly the lowest possible; out of 50 reviewers, 76 percent gave one star. It appears, at least, that some of those consumers actually got the product, but reviews on other sites showed people complaining they had yet to receive their OxyBreath Pro.

RiskIQ dug into some of the websites that had the words “OxyBreath Pro” directly in their domain names, and found they were recently created, according to Herman. “Timing of domain registrations could indicate this product was created to take advantage of the pandemic,” Herman says.

Join the club

Of course, the shadowy network of websites behind these particular masks is only a small piece of the problem for platforms like Facebook, Google, Amazon and Snapchat. These multibillion-dollar companies deploy automated advertising infrastructure that allows marketers to open accounts and start placing ads on their sites, and they have third-party ad networks that deliver ads to other web properties.

The internet companies police ads mostly through machine learning and artificial intelligence, teaching their filters to detect images, analyze keywords and scan website links for bad actors. Facebook and Google have beefed up their security protocols in recent years, especially following the 2016 election when they uncovered networks that use many of the same tactics as these coronavirus marketers. In 2016, overseas operators built websites trading in disinformation, now they’re using those same tools to capitalize on the latest tragedy.

Since January, when coronavirus first surfaced widely in China, the platforms have been inundated with this new breed of opportunism. In February, Facebook said it would crack down on ads that “refer to the coronavirus and create a sense of urgency, like implying a limited supply, or guaranteeing a cure or prevention.” The social network has specifically targeted face masks marketers, too.

At the start of March, Amazon announced it removed a million products related to coronavirus from its retail site, which was dealing with underhanded sellers marketing products with specious claims and also gouging prices on necessities.

Facebook and Amazon did not return requests to comment for this story.

A Google spokeswoman said the company has been removing more than a million ads a day related to COVID-19, and most of those are for face masks.

“We’ve blocked ads for products that aim to capitalize on coronavirus, including a temporary ban on face mask ads,” the spokeswoman said in an email statement. “We’ve seen highly opportunistic advertisers try to run an unprecedented number of these ads on our platforms. We have a dedicated task force working to combat this issue and are removing millions of ads per day. In this instance, we removed the ads quickly after they were flagged to us. We’re monitoring the situation closely and continue to make real-time adjustments to protect our users.”

Avoiding detection

After being notified of the activity, Google’s spokeswoman said the company uncovered hundreds of ads coming from thousands of accounts related to the OxyBreath Pro mask. All of which have now been shut down.

The face mask marketer, however, shows how evasion tactics have evolved even as the platforms’ detection systems grows more sophisticated. On Google, OxyBreath used search ads, which don’t have images, so they avoided image recognition software that spots face masks. The marketer also didn’t use “face masks” or “coronavirus” terms in the descriptions, avoiding keyword filters.

Facebook is dealing with a similar predicament. The company has been removing these types of ads for weeks, yet new ones seem to slip through. In the middle of March, Rob Leathern, Facebook’s director of business integrity, tweeted about how the company was still refining its filters to nab bad ads. “We have been preparing for more automation catching these ads but it is not there yet,” Leathern said at the time.

On Thursday, Facebook even sued a software maker that created a program to help fraudulent websites evade ad filters.

At Twitter, a spokeswoman there said that it had not found any ads for OxyBreath Pro masks, but there were multiple accounts creating spam posts about the product. Twitter shut those down after being alerted.

On Snapchat, too, there were ads for the product, which could be a surprise to many in the ad community because the app has been largely sheltered from many of the abuses that plague its rivals. Snapchat is more of a closed app, which makes it more difficult for bad actors to garner attention, unless of course they buy ads to show their content to people.

On Snapchat, the personal video messages created by users are typically shared privately, so an organic account won’t have as many tools to reach a wide audience as it might on Twitter or Facebook. Also, Snapchat’s media section, called Discover, vets publishers to keep a lid on fake news.

Snapchat has promoted Discover as a media offering that can combat disinformation, the kind that proliferates on Facebook, Twitter and YouTube. It’s true that Snapchat is more shielded from some of the worst behavior but it is clearly not free from it.

Snapchat ads for OxyBreath Pro linked to a site called “simple discount finder,” which was also registered to Novads OU.

When contacted about the ad, a Snapchat spokesman said that the company had recently implemented a ban on promoting items like personal protective equipment, which covers face masks. This advertiser would have been in violation of that new policy.

After Snapchat removed the advertiser, the website for “simple discount finder” disappeared from the internet.

COVID ad conundrum

It’s not so simple for the internet companies to craft policies around coronavirus ads, and the rules have been shifting as the crisis continues. With the pandemic lasting months, it has become nearly impossible for platforms to dictate to brands that they can’t mention the issue.

At the start of the outbreak, Twitter and Google had taken a hard line against discussing the virus in ads. They drew a sharp line that was meant to clamp down on any illicit advertisers, but it also affected major brands that were just trying to put out messages related to their efforts during the pandemic.

At the beginning of this month, Twitter eased up on its policy and began allowing major brands to craft messages related to the pandemic. Advertisers that use Twitter’s automated ad system still can’t run COVID-19-related promotions, but ones with a direct relationship to Twitter’s sales team can. Twitter highlighted campaigns from Uber, UPS and Starbucks to show how to advertise appropriately around the subject.

Google has said it is still working on its coronavirus ads policy to allow certain advertisers to market their products and services related to coronavirus. Meanwhile, Snapchat allows brands to mention coronavirus in their ads, given how central the topic is to the daily conversation.

Snapchat has banned, like the other platforms, the promotion of personal protective equipment like face masks. However, even that could cause problems for legitimate brands that want to sell products that are in high-demand. The U.S. Centers for Disease Control and Prevention has issued guidelines that people should wear face masks when out in public. Where are people supposed to buy those masks when they become readily available?

At some point, it would make sense to allow brands to promote their face masks in internet ads, says Brian Wieser, Group M’s global president of business intelligence. “The sensitivity will vary as time progresses,” Wieser says. Face mask advertising from, say, luxury brands could be permissible, “If we go into a point of time where masks are just deemed to be something that’s normal, common social behavior,” Wieser says.

Until then, the platforms will be on the lookout for the worst abuses of their services, which is something that federal authorities in the U.S. have been warning about. The Federal Trade Commission has issued alerts to consumers about marketers promoting unproven medical claims in ads. The U.S. Department of Justice has said it is watching out for fraud that could flourish around the pandemic.

Carl Settlemyer, staff attorney at the FTC’s Bureau of Consumer Protection, says that the internet platforms are the first line of defense to crack down on digital scams. If the internet companies show they are making their best effort to thwart fraudsters, then the federal government would typically not take action against Facebook, Google, Snapchat and others for bad actors that slip through the cracks.

“We expect that they will live up to their commitment to try to police their platforms for any advertising that would be potentially harmful to the public, including false deceptive advertising,” Settlemyer says.