Today in Digital Marketing - The Lawsuit That Could Break Marketing AI

Starting point is 00:00:00 It's Monday, November 21st. Today, reduced engagement, serious brand safety issues, why one brand spending millions on Twitter ads finally bailed out. Also, who's liable when marketing AI makes a mistake? Snapchat struggles to attract influencers, Amazon's getting creepier, and in case you think your day is bad, just be glad you're not the social media manager at KFC. I'm Steph Gunn filling in for Todd Maffin. Here's what you missed today in digital marketing. While big advertisers bailing out of Twitter have garnered headlines, one interesting blog post now gives us an interesting inside look at the specific reasons media buyers are uncomfortable.

Starting point is 00:00:41 Yesterday on the anonymous website Blind, one media buyer posted about why they paused their $750,000 per month Twitter ad campaigns. They said they're a director at a medium-sized B2B tech company with a team that spends roughly $80 million annually on advertising, adding that Twitter had accounted for about a tenth of their media mix. Apparently, they tried to hang in there and kept their campaigns live for two weeks after the takeover, in the belief that efficiency would improve with fewer advertisers. But the director says they've since found they were wrong. Here's what they found, quoting from the post. Performance fell significantly. CPMs didn't drop, but our engagement went way down.

Starting point is 00:01:22 Maybe it's a shift in users on the platform. Maybe it's ad serving related. Serious brand safety issues. Our organic social and customer service teams got dozens of screenshots of our ads next to awful content. Replies to our posts with hardcore anti-Semitism and adult spam remained up for days, even when flagged. Our entire account team turned over multiple times in two weeks. We had multiple people supporting our account and they all vanished without so much as an email. We finally got an email with a name for an account manager last week, but they quit and we don't have a new one yet. And finally, the ads user interface is very buggy and

Starting point is 00:02:05 two-factor authentication broken. One of my campaign managers logged in last week and found all our paused creatives from the past six years had been reactivated. Campaign changes don't save. These things cost us real money. Marketers are closely watching a lawsuit which could shake up the burgeoning copywriting ai space a lawsuit filed earlier this month targets github copilot a tool that automatically writes working code when a programmer begins typing however some coders notice that copilot occasionally reproduces code snippets that are taken from the millions of lines available in public code repositories. The suit argues that GitHub is infringing copyright because it does not include attribution when Copilot reproduces open source code covered by a license requiring it. Coders

Starting point is 00:02:57 have, of course, always studied, learned from, and copied each other's codes. However, not everyone agrees that AI should do the same. Legal experts suggest this lawsuit may affect the broader trend of generative AI tools, like the kind marketers have begun to rely on. The algorithms used in AI programs to create headlines, marketing copy, photographs, and illustrations are all trained on existing work by humans. And this could affect the big players too, like Meta, which has prioritized a more automated approach to ad creative. There are big, unanswered questions in this. Who is liable if Meta's adbots write marketing copy that was largely inspired by another brand's work? The advertiser? Meta? Anybody? GitHub's CEO noted

Starting point is 00:03:44 that Copilot now comes with a feature that prevents copying existing code. However, whether this provides adequate legal protection remains to be seen. Even with such legal issues in the air, marketers continue to invest in AI technology like natural language processing and data-driven personalization. A new study from Digiday published today looks at how brands have changed how they use artificial intelligence in marketing and customer service. Apparently, marketers mainly use NLP in the form of chatbots to streamline service responses and increase sales. Other forms of NLP like social listening

Starting point is 00:04:21 continue to be popular among marketers who need to track brand mentions to make business and product decisions. The big use of AI, though, has been in data-driven personalization. More than two-thirds of respondents reported that it is a top priority this year. As for AI developing creative, so far the focus has been on targeting individual users for a hyper-personalized experience. The top consumer data points marketers use to fuel data-driven personalization continue to be things like site behavior, location, and search history. Finally, marketers who do not currently use data-driven personalization are evenly divided

Starting point is 00:05:00 on whether to invest in it in the future, with half of respondents saying they will and the other half saying it is not relevant to their business. AI influences more than marketing tasks, of course. It's even being used as a kind of HR nanny system at the world's largest e-commerce platform, Amazon. The Verge reports today about the impact being felt by an offshore workforce training Amazon's warehouse monitoring algorithms. Reports say warehouse workers stocking Black Friday deals are monitored by three cameras pointed at them. Amazon then pays workers in India and Costa Rica to watch hours of endless warehouse video streams for just hundreds of dollars a month.

Starting point is 00:05:45 Video reviewers said their primary role was stock management, but they can also record errors made by their colleagues. 2X employees noted that reviewers could raise stow etiquette issues if they saw someone breaking Amazon's rules on camera. The AI camera system also monitors the stowers' movements, and if they're underperforming, a video is sent to someone thousands of miles away whose input helps to improve Amazon's machine learning tools. Ultimately, says Amazon, this work improves the computer vision system,

Starting point is 00:06:17 which learns from their responses and becomes more accurate. According to the report, reviewers have to analyze up to 8,000 videos a day, with their output ranked against that of their colleagues. Several workers reported their health deteriorated as they faced extreme pressure to meet a metric of 99% accuracy. Amazon denied a number of the allegations in the Verge's story, saying the piece paints a misleading picture and that it does not represent the vast majority of its team. Do you have business insurance? If not, how would you pay to recover from a cyber attack, fire damage, theft, or a lawsuit? No business or profession

Starting point is 00:07:00 is risk-free. Without insurance, your assets are at risk from major financial losses, data breaches, and natural disasters. Get customized coverage today starting at $19 per month at zensurance.com. Be protected. Be Zen. Like many other platforms, Snapchat has been trying desperately to lure creators to the platform, hoping that would drive increased traffic. But it doesn't seem to be working. Why? A piece up on Digiday today suggests it largely comes down to money. One creator explained that while Snapchat was never a priority for him, he did value its daily $1 million Spotlight creator pot in the early days.

Starting point is 00:07:41 But just nine months after launching, creators were leaving Snapchat after payments dried up and the $1 million per day pot decreased to millions per week, which eventually became millions per year. Money, however, is not the only challenge creators face. Some are still struggling to feel part of the Snapchat family. For example, another influencer spoken to for the study believes the company can still improve its relationships between reps and creators. It was easy to get lost in the mix and difficult to build one-on-one relationships. Finally, discoverability remains a challenge. While TikTok's algorithm has been praised for

Starting point is 00:08:21 its ability to surface compelling videos, identifying content that would do well on Snapchat is a much more difficult task. On other platforms, one creator explained he has a general idea of how each video will perform, which makes creating sponsored content easier. His videos' views typically average in the hundreds of thousands, but on Snapchat, it's not uncommon to get only a thousand views. Here's yet another example of how not reviewing your brand's content before it goes out can be disastrous. KFC has apologized for pushing out a promotional notification on its app, telling German customers to treat yourself on Kristallnacht. If you're not familiar with

Starting point is 00:09:06 Kristallnacht, it's far from festive. Instead, it references the attacks in 1938 that killed 90 people and destroyed Jewish-owned businesses and synagogues. It's regarded by many historians as the beginning of the Holocaust. So what happened? KFC said they were using an automated content creation process that relied on a calendar that included national observances. And nobody apparently reviewed what a very basic AI bot wrote and published. KFC has apologized and said it will review its procedures. I've worked for this podcast for over a year now so I feel like this is a safe enough place to admit that I am a Swifty and I was part of the 7 billion people that tried to get Taylor Swift tickets last week so I waited for five hours when the tour was first announced to register for the presale. That wasn't even to get tickets.

Starting point is 00:10:09 And what happened? I got waitlisted. So I was not able to get tickets. And if you did get them last week, one, I hate you. Two, do you need a plus one? Thanks for listening. I'm Steph Gunn. Todd's back from vacation tomorrow.

Starting point is 00:10:41 It's the season for new styles and you love to shop for jackets and boots. So when you do, always make sure you get cash back from Rakuten. And it's not just clothing and shoes. You can get cash back from over 750 stores on electronics, holiday travel, home decor, and more. It's super easy. And before you buy anything, always go to Rakuten first. Join free at rakuten.ca. Start shopping and get your cash back sent to you by check or PayPal.

Starting point is 00:11:10 Get the Rakuten app or join at rakuten.ca. R-A-K-U-T-E-N dot C-A.

Today in Digital Marketing - The Lawsuit That Could Break Marketing AI

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.