AI: Ready, Set, Regulate & Innovate!

What Happened: Early reviews are in, and the Googs still has some work to do. We made a joke last month about Magi stealing your content; well, we weren’t far off. First, “Project Magi” is different from the Search Generative Experience (SGE) that Google is currently testing. And second, Twitter has been lighting up with complaints about SGE citing complaints ranging from missing and unclear attribution of content to privacy complaints and search console requests. This is just the first of a long line of ticket items as Google launched the Beta through Search Labs and continues to note that “it’s the early days” and they are still “experimenting” and encourage “feedback.”

One user asked about personalization, to which Google responded that they “don’t believe” they are, but SGE responses can change based on the context of what was previously entered. This confusion is likely due to Project Magi, which was touted as “a more personalized Google Search” pre-SGE launch.

Additional chatter around Google SGE highlighted that the experience is too slow. The question is, why would I wait for the SGE response when I can just scroll down? Google wasted no time in fixing this issue, on June 8th a number of quality updates and “major improvements” to cut the generation time in half when launched. Barry Schwartz said, “now just really fast. It is not as fast as a search result set being returned, but it is faster than other AI search engines. And yes, it will get faster.”

SGE isn’t the only “reimagined” search experience at play right now. Amazon is planning on reimaging its search with a conversational AI chatbot. The AI is supposed to answer product questions, perform comparisons, and make recommendations. All features and use cases we see the new SGE experience used for too, though it’s specifically been cited as a response to OpenAI’s tool, ChatGPT. Amazon’s answer to ChatGPT has caused some head turns as the tech has given the impression that it’s “incomplete.” Users were given access but interestingly told not to test it.

Users aren’t blindly bought into AI either. Social is a good indication of user engagement and predicted behavior, and everyone hated the Snapchat bot. In leaked images, it seems like Meta is considering adding a chatbot to Instagram, though the implementation is challenging and fraught with user safety concerns especially given Snapchat’s My AI bot propagated “harmful and inappropriate advice.” Aside from the need for safeguards so adverse engagement is avoided, chatbots have the ability to open the doors to a new level of personalization and engagement in ad targeting.

Regulation is still a topic of debate. Senate hearings on AI regulation are still underway, which have been decidedly bipartisan and collaborative, though the path forward still isn’t clear. There are a few different regulatory options on the table, large companies are promoting licenses to make big models, though smaller companies and start-ups feel this is just a gatekeeping tactic to keep the barrier to entry too high for smaller competition.

An industry group spearheaded by Adobe called the Content Authenticity Initiative (CAI) which is aimed at bringing transparency to AI-generated content in a “nutrition label” format will help put the industry ahead of any regulation. With the growing conversation becoming more complex as the lines between tech and business continue to blur, a clear path forward is still elusive.

Adoption is booming, no surprise there, though rather than creation major brands are leveraging “plug-ins.” Microsoft announced it will leverage the same plug-in standards as OpenAI, meaning they can work across ChatGPT, Bing Chat, and Microsoft 365 Copilot. ChatGPT plug-ins are only available for select businesses but have been likened to “day one of the Apple App Store” and said to have the ability to “transform how consumers interact with brands online.” So, get on that.

On a more dramatic note, we like those, a “bunch of celebrity AI experts” signed a 22-word statement (warning) about the existential risks of AI. We’ll save you the click-through: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war.” Ya know what’s interesting about those signees? Who they are: Sam Altman. Demis Hassabis. Geoffrey Hinton. Stuart Russell.

Cats outta the bag now, though. (… *cough cough*… ^their cat, ^their bag)

Impact on Search: eMarketer hosted a virtual summit featuring Lily Ray and other industry leaders to discuss what’s redefining search advertising today. The three main points included changing consumer behaviors, an influx of digital channels, and, of course, generative AI. Younger generations are turning to TikTok and YouTube Shorts to initiate their search which engages in personalities over brands, Google is keen to this shift as the launch of “perspectives,” a search filter aimed at providing content from personalities across the web, in an effort to stay in the path of first searched. With this, diversification is key to successful marketing strategies today. Though paid search is fundamental, avenues such as social media, video, and other mobile channels must be present. And, coincidentally, Ad spend has been down for 10 straight months. Though declines may be attributed to the industry contraction post-pandemic boom and economic strains, there is still growth just not near as much as 2022’s high rate.

Last but not least, Generative AI is still firmly in the grey area, especially for highly regulated industries, like healthcare and finance, and there are some “existential” questions about how the information used is sourced.

Attribution concerns are a major sticking point for most in the industry, especially news publishers. Digital Content Next (DNC), a trade association including large publishers like The New York Times, Bloomberg, and Axel Springer; warned that generative AI tools may violate copyright laws. The guidelines DNC drafted highlight “AI's use of publishers' content often exceeds fair use under copyright and general law.” Major concerns included generative AI tools’ access to paywall information, disrupting site traffic, and accessing content archives without proper attribution or compensation.

Generative AI is the rogue wave we’ve all known about for years, but “suddenly” it’s heading straight for us. It’s unclear who’ll be riding that wave when it comes crashing down, but as far as tech developers go, it’s who’s leading the pack and countless others are entering the market daily. As this industry continues to shake out, one thing we’ve learned over the years is that users’ behaviors will be impacted, and we are already seeing how users are shifting where they start their search online.