Companies investing in AI are increasingly investing more to market it.

At least $40 million has been spent this year on advertising AI products across print, digital, TV and events, according the ad-tracking firm MediaRadar. And in recent months, momentum is picking up.

Tech giants and startups racing to stand out spent more than $35 million in just the third quarter of 2023 — a 550% increase over the third quarter of 2022 — noted MediaRadar, which analyzed spending data through September for more than 360 advertisers. The total number of third-quarter advertisers also increased by 135%, jumping from 80 in 2022 to more than 180 in 2023.

Only 6% of AI advertisers spent over $100,000, contributing $36 million to total spending in the first nine months. The remaining 94% collectively spent just $4 million. MediaRadar highlighted new major advertisers like IBM and Salesforce — for products such as WatsonX and Slack GPT — along with startups like cloud communications provider Dialpad.

The findings illustrate what might be in store for the category as AI expands into various business and societal sectors. (In September alone, overall AI advertising skyrocketed 730% from August, which MediaRadar said was thanks to new advertisers and a spending surge from IBM across TV and online video.)

In terms of AI advertising formats, 34% of total spending went to TV, 29% to online video, and 12% to print. Digital advertising made up $19.8 million or 49% of the total spend. This included $11.7 million for online video, with digital display and paid social each receiving around $3 million. Event exhibits and sponsorships accounted for $1.8 million, or approximately 5%.

As tech giants and startups both collaborate and compete, it’ll be worth watching how marketing strategies evolve across various ad formats. Can newcomers create compelling campaigns as a way to stand out, or will current incumbents use their size and familiarity to stay on top?

Earnings season: A glimpse at AI in Q3
It’s once again earnings season, and AI is being mentioned plenty in Q3 results and accompanying earnings calls.

As Meta increasingly focuses on generative AI, CEO Mark Zuckerberg expects AI will be “the largest area of increased investment.” (Meta’s Llama large language models were downloaded more than 30 million times in September.) However, he said it’s hard to predict the mix of people’s chats with bots versus chats with humans across various apps. “I do think that the fundamental technology around generative AI is going to transform meaningfully how people use each of the different apps that we build,” Zuckerberg said. “I think for the Feed apps, I think that over time, more of the content that people consume is going to be either generated or edited by AI.”
Sundar Pichai, CEO of Alphabet and Google, said the company will continue working on its generative search products like Bard and also will experiment more with new ad formats native to its search generative experience (SGE). He also mentioned the total active generative AI projects using Google’s Vertex AI platform has increased by 7 times. When an analyst asked if search activity will remain in the search bar or become decentralized to Bard and elsewhere, Pichai said people have have “always looked for [information] in many, many different ways” and that he doesn’t think that’ll change.
Snap revealed that more than 200 million people have used its My AI chatbot while total sent messages have surpassed 20 billion. Most people are still using it on a weekly basis rather than daily, said CEO and co-founder Evan Spiegel. When asked about the performance of ads within My AI — which are powered by a new partnership with Microsoft — Spiegel said the early click through rates “are very healthy” when ads are relevant to users, but didn’t mention any particular numbers. “The big focus has really been on improving relevance, continuing to drive relevance for Sponsored Links in My AI,” Spiegel said. “So far, in terms of signal integration, right now, we’re using it to inform interest categories. I think there’s a lot more we can do there. But we’ve sort of taken that initial step again with the test-and-learn perspective.”

During IBM’s earnings call, Chairman and CEO Arvind Krishna spent more than half of his opening remarks on the topic of generative AI. Along with mentioning IBM’s various AI models, Krishna said the top three enterprise AI uses based on current customer interest will be code modernization, customer service and productivity tools. IBM also plans to debut its new platform by the end of the year to help companies mitigate risks.
For Microsoft, CEO Satya Nadella provided a range of updates related to its gen AI tools. On the company’s earnings call, he said Microsoft Bing users have engaged in more than 1.9 billion chats and generated more than 1.8 billion images. In the past three months, search and news ad revenue increased 10% and 9% year-over-year — excluding traffic acquisition costs — while Microsoft Edge has now “gained share for 10 consecutive quarters.” Nadella also mentioned that total LinkedIn members watching AI-related courses increased 80% in the past quarter.

Beyond the black box: Regulatory updates

The Frontier Model Forum — an industry organization focused on developing safe and responsible AI — announced Chris Meserole as its first executive director. (Meserole was previously director of Brookings Institution’s Artificial Intelligence and Emerging Technology Initiative.) The Frontier Model Forum’s founding companies — Google, Microsoft, OpenAI and Anthropic — also announced a $10 million investment for a new “AI Safety Fund.”
On the regulatory front, nearly two dozen tech execs, researchers, investors and other experts visited Capitol Hill for the second “AI Insights Forum,” a closed-door meeting hosted by U.S. Senator Chuck Schumer.

The United Nations announced a new AI advisory board to explore various opportunities and risks and also focus on global governance related to AI.
Ahead of this week’s AI Safety Summit, the British government announced plans to create an “AI Safety Institute.” It also published a report examining processes for safe AI. According to the accompanying policy paper, 62% of British residents said the government should test AI to make sure it’s safe while results in other countries varied.

This week, the International Association of Privacy Professionals (IAPP) will hold its 2023 AI Governance Global conference in Boston. Speakers include former New Zealand prime minister Jacinda Ardern along with various tech execs and privacy experts.

Prompted products

Amazon debuted a new AI product image generator for advertisers. In one test, placing a toaster on a kitchen counter next to a croissant led to 40% higher click-through-rates than an image of the same toaster with a standard blank white background.

In other image-related news, Shuttershock debuted a new tool that lets people use AI to edit real photos.

A new tool called Nightshade lets artists “poison” AI models with corrupted training data. The tool, first reported by MIT Tech Review, gives creatives a way to fight back against AI models amidst a range of copyright concerns. (Read more about it in the research paper published by the University of Chicago’s computer science department.)

Qualcomm debuted its next generation Snapdragon chip, which will power AI platforms on various mobile devices. It also announced the chips will include technology from TruePic that will help improve content authenticity at the device level.

Lenovo and Nvidia announced a new partnership to help enterprise companies quickly adopt generative AI.

Other AI news from Digiday:

Forbes debuted Adelaide — a new generative AI search tool, that lets visitors ask questions in the website’s search bar. (Digiday)

The NBA is experimenting with generative AI content for the 2023 season while also testing new ways to analyze and categorize game play. (Digiday)

As generative AI grows on social media, agencies and brands are exploring new ways to measure content. (Digiday)

For Digiday’s ongoing WTF series, Tim Peterson created a video that explains content credentials and why they’re important in the era of generative AI. (Digiday)