ExchangeWire’s Post

Google has published its annual Ads Safety Report highlighting the actions we took last year to keep users and advertisers safe on our platforms and to maintain a healthy digital advertising ecosystem. Generative #AI tools have enabled us to take a major leap in our enforcement efforts, specifically Large Language Models (LLMs), so that we can better keep people safe online. At a high-level, in 2023: ➡ We blocked or removed over 5.5 billion ads for violating our policies. That's more than 9,000 ads per minute. ➡  We suspended more than 12.7 million advertiser accounts for egregious policy violations, nearly double the amount we suspended in 2022. ➡  We removed ads from over 2.1 billion pages last year. ➡  We added or updated 31 policies for both advertisers and publishers. Read the blog post for more details on how we’re keeping the ads ecosystem safe for advertisers, publishers and users. goo.gle/ads-safety-report #AdsSafetyReport

At Google, there’s nothing more important than keeping you safe online. That’s why we work around the clock to develop advanced technologies to better safeguard all of our products, including the ever-shifting digital advertising ecosystem. Releasing today, our 2023 Ads Safety Report details the past year’s efforts to prevent malicious use of our ads platforms. Most notably, our teams have embraced generative AI and specifically Large Language Models (LLMs) to rapidly review and interpret content at a high volume while also capturing important nuances within that content. Let's take a look at some key statistics from the report: 📌 Blocked or removed over 5.5 billion ads for violating our policies. (That's more than 9,000 ads per minute.) 📌 Suspended more than 12.7 million advertiser accounts for egregious policy violations, nearly double the amount vs 2022. 📌 Blocked or restricted ads from over 2.1 billion pages last year. These achievements mark only the beginning of our efforts. With scams, fraud and bad ads on the rise, we will continue to develop new policies, strengthen our enforcement systems, and leverage the power of generative AI and LLMs to keep you safe and to maintain a healthy ad-supported internet. Learn more about our latest Ads Safety Report: https://lnkd.in/g8pAzWAp

Google's AI Overview summaries will feature in the search results of over one billion Google search engine users by the end of the year. It has potentially major implications for the flow of traffic to websites, and how website owners have their content summarised. With an advertising business in New Zealand approaching $1 billion in value, Google will be making sure its new AI features are revenue-neutral at worst, and vastly more profitable at best. But what about content creators trying to find an audience? Marketers are having to adapt their approach to deal with AI-driven search summaries. My BusinessDesk NZ column (premium) https://lnkd.in/dAQXXBZ4

📣 SEVEN marketing stories you should probably know this week Industry News 📣 - OpenAI will pay to license content from The Financial Times for ChatGPT - including FT article summaries, links to articles and usage for training data (The FT - Official) - Bytedance would rather shut down TikTok in the U.S. than sell to protect algorithm. Instagam & snap could be the big winners (Reuters) X's new video app is coming to smart TVs. The layout resembles YouTube more than a little! (X - Official) - Q1 earnings point to a growing digital advertising market. Ad revenues: Google's 13%, YouTube 21%. LinkedIn 10%, Meta 27%. Snap 21%. Google CEO says AI overviews are increasing Search usage. (Google - Official) - Google delays third-party cookie phase-out (again) until 2025 (maybe?) due to concerns from UK regulators about potential competition issues. (Google - Official) - Amazon Prime & YouTube look to acquire rights for NBA games. (USA Today

I think it's completely valid to assume new generative-AI systems will make it easier for some coders to deploy malicious programs, and there's already been some interesting research (BlackMamba @ https://lnkd.in/gSfhUTC6) from HYAS into polymorphic malware using AI to dynamically obfuscate C2 connections and in-memory loading schemes. In the ad tech sector, domain/app/channel transparency is essentially required to actually cash out a scheme, so some aspects of how generative AI can generate new forms of ad fraud require slightly more nuance than just "will generative AI make it easier to dynamically change C2/redirectors?" The industry will absolutely see types of advanced machine-written code within certain data flows, but this will likely be "easier to spot" (maybe not easier to reverse!) than AI-written content filling up content farms, subtly not cloned or copied, but nearly impossible to differentiate from original content beyond the rhetorical flourishes currently seen in AI-generated writing. In the future, how will your ad buyers know whether they are buying on a publisher with content written by humans, compared to a publisher with content written by machines? And will your ad buyers try to completely avoid specific types of content or will your inventory slowly fill with publisher sites composed mostly of AI-generated content? I provided some additional thoughts about these challenges in last week's AdMonsters eletter @ https://lnkd.in/g3nXDyNa “Ad buyers navigating these placement challenges will need to make smart buying decisions across all types of web properties and media channels while also communicating their content preferences when serving their ads. In the coming months and years, we will see most buying platforms having customers asking for a way to avoid buying ads on AI-written content, or at least give the ability to know how many of their impressions served on this content and the demands for this are already showing from the big media groups and most sophisticated ad buyers.." Check out their piece (https://lnkd.in/g3nXDyNa) for some additional thoughts from industry experts, and please share any thoughts you have about how you think generative AI will impact the ad tech industry. We need to be thinking about this sooner than later! 🖖

Don’t let AI disguise a backwards slide for agencies and brands 🤖 Resharing Artiom Enkov's piece in The Drum, including: ➡️ Why Meta and Google's Advantage and Performance Max take us back to the origins of programmatic. ➡️ ‘Dumbing down’ offerings for SME advertisers, at the expense of bigger clients? ➡️ Allowing "Google to get closer to brands, removing agencies from the equation" - even as tech giants cut their own staff, in some cases automating customer support. Read in full ⬇️

Discover the latest insights and trends from #AdWeek 2023, including discussions on privacy, AI's role in advertising, and the importance of first-party data, as shared by Twilio Segment and industry leaders. https://lnkd.in/eX4Kkb7q

What can companies diving deep into generative AI learn from the mistakes ad tech made in the past? Lots, and if you want to know, read Marty Swant's excellent piece in Digiday.

Just finished reading some insights from Ad Week 2023, courtesy of Segment. Here's a quick rundown: 1. Personalization is no longer a nice-to-have, it's a must-have. Tailored ads engage customers and drive conversions. 🎯 2. First-party data is the new gold. As third-party cookies phase out, businesses need to leverage their own data for deeper customer insights. 🍪➡️💡 3. Customer Data Platforms (CDPs) are the new norm. They help businesses collect, unify, and activate customer data for personalized experiences. 🔄 4. AI is revolutionizing advertising. Predicting user behavior and personalizing ad experiences is the future. 🤖 5. Privacy is paramount. Businesses need to respect customer data privacy while delivering personalized experiences. 🔒 6. Collaboration is key. Aligning marketing, product, and engineering teams ensures everyone is working towards the same goals. 🤝 In a nutshell, the future of advertising is all about personalization, leveraging first-party data, using CDPs, and respecting customer privacy. Exciting times ahead! 🚀 #AdWeek2023 #Personalization #FirstPartyData #AI #Privacy #TwilioSegment

A bipartisan group of senators has introduced the Content Origin Protection and Integrity from Edited and Deepfaked Media Act to protect artists, songwriters, and journalists from having their content used to train AI models without consent. Authored by Senators Maria Cantwell, Martin Heinrich, and Marsha Blackburn, the bill mandates that AI developers include provenance information on digital content, preventing its use in AI training and enabling creators to sue for unauthorised use. The US National Institute of Standards and Technology will set guidelines for detecting AI-generated content. This initiative comes amidst a surge in AI regulation efforts, aiming to increase transparency and combat harmful deepfakes. https://lnkd.in/dtVyGjjb