Top Stories On AI Accuracy, Business Adoption, and More
Groundbreaking research on AI accuracy and what to watch out for. Google Ad tools to trial. Unprecedented stats on AI adoption. Let’s dive in.
Empathetic Chatbots are More Likely to Lie to You
TLDR: A recent study from Oxford University found that LLMs tuned to anticipate users’ feelings were more likely to produce inaccurate results. Researchers compared results from LLMs tuned for higher emotional “warmth” with those from LLMs that received no fine-tuning. The “warm” models were over 60% less accurate.
Why this matters: You may remember when ChatGPT became a complete sycophant a few months ago after OpenAI pushed an update. While that was an extreme case, this phenomenon is still worth watching out for. Much like humans, LLMs are often trained to “soften the blow,” so to speak, when delivering information that a user won’t like. That business idea you had that your favorite chatbot said was genius? It may be holding out on you.
Source: Ars Technica
Get Ready for New Google Ad Tools
TLDR: Google just launched new capabilities in AI Max, its suite of AI-powered tools within Google Ads. This includes AI Brief, which automates previously manual settings like audience guidelines, messaging parameters, and search targeting.
Why this matters: While AI Max already offered a number of AI-powered optimization tools, Google’s latest rollout promises to eliminate some persisting guesswork in ad campaigns. The only way to see their true impact? Testing it out. Marketing teams should start trialing use cases and seeing what works.
Source: Marketing Brew
AI Reaches 53% Population Adoption. Rapid Growth is Outpacing Governance.
TLDR: Stanford’s 2026 AI Index Report reveals AI is the fastest-adopted digital technology, taking just 3 years to reach 53% population adoption. The internet, for reference, took 7 years to reach majority global population adoption. At the same time, only a third of Americans feel confident about AI regulations. The public is optimistic about AI advancement and confused about AI guardrails.
Why this matters: We need strong AI ethics on the individual level, but it shouldn’t stop there. People want clear guidelines for responsible use in the workplace, as consumers, and in public programs like education and administration.
Source: Stanford 2026 AI Index Report
Since large-scale AI governance won’t come overnight, enacting a clear AI action plan within your own organization is a positive place to start.
We’re Working on a System that Grows Online Visibility.
InboundAV is fine-tuning a system to identify low-hanging fruit when optimizing businesses for online visibility. The idea: maximize the probability of your website getting seen and cited by answer engines, including Google and major LLMs, without a complete overhaul of your web content. Wanna learn more?
