📫 Subscribe | 🗣️ Hire us to speak | 🎓 Online Courses | 📅 Events

Rapidly expanding AI data centers have emerged as a major flashpoint across the political spectrum. MAGA Republicans fear AI-powered surveillance, environmentalists cite carbon emissions, and rural residents raise concerns about heavy water and power consumption. Local communities and voters across key mid-decade contests are increasingly opposing tech industry tax incentives that support these massive facilities. The election outcomes will offer a decisive test on whether public pushback will reshape state-level tax subsidies and restrict future data center infrastructure development.

Right after today’s midterm elections, OpenAI plans a policy and PR push aimed at a Washington where Democrats will likely win the House, and at the 2028 presidential candidates. The plan includes proposals for mandatory “speed limits” on the most powerful AI systems, plus ideas for helping workers through disruption, expanding AI training, and shaping global AI standards. On data centers, OpenAI will offer community compacts, propose state laws giving residents a real say in projects, and push to rebrand its sites as “AI compute campuses,” pointing to a Pike County, Ohio, project that promises thousands of jobs. Policy chief Chris Lehane frames it all around keeping people in control of AI and calls the midterms a preview of 2028, when he says AI will be the central issue.

Google introduced Gemini 4 Carbon, capturing the top spot on updated benchmark evaluations with state-of-the-art cybersecurity capabilities and long-horizon task performance. Building closely on Gemini 4 Argon, Carbon delivers substantial improvements in vulnerability discovery and autonomous software engineering. However, public access remains strictly limited as Google initially rolls out the model to selected cybersecurity defenders and trusted partners before a broader release.

Lean TECHniques works alongside Midwest companies to leverage technology for business impact. Through advisory, coaching, and embedded delivery, Lean TECHniques helps organizations develop high-performing teams, create environments that fuel innovation, and deliver best-in-class customer experiences. Core areas of expertise include software, cloud, artificial intelligence, product management, and technology leadership.

Canva launched Brush – its first vision-language model for design, built to edit designs instead of making new ones from scratch. Tell it what to change, and it can restyle an image, make a change across a multi-page design, or remix assets for different channels – with text, imagery and layout staying fully editable afterward. Canva says it’s up to 69% faster for general design editing tasks and 46% cheaper per edit than its previous design agent, and Pro subscribers can try it now by selecting ‘Brush’ from the dropdown list in Canva AI. Editing is where most of us actually spend our time in Canva, so this may end up mattering more than another image generator.

This was a fun one. Snider turned a photo of him sitting at his desk (see the original image here) into the image below.

A new study of 55,198 AI-generated answers found that 57% of links in Google AI Mode product answers led to Google Shopping product pages, and AI Mode linked only to Google’s own pages in 46.5% of U.S. product answers. Findrix tested 18,400 buyer questions across 35 U.S. markets in mid-October, and the findings measure links shown, not actual clicks or purchases. The takeaway for brands: you can show up in AI search without getting the visit, so the product data you give Google matters as much as your website.

Prompt Engineering: The art and science of structuring instructions, context, and examples so an AI model gives you the exact output you need. Think of it less like writing a software script and more like briefing a talented human assistant who has total knowledge of the subject matter but zero context on what you want right now.

Good prompt engineering isn't about using magic keywords or secret syntax. It’s about clarity: giving the model a defined role, clear constraints, relevant background information, and an example of the desired format. If you ask an LLM to "write a marketing email," you get a generic placeholder. If you tell it, "Act as a B2B SaaS copywriter, write a three-paragraph launch email for a new reporting feature, and keep the tone conversational without using buzzwords," you get something you can actually use. As AI models grow more capable, raw prompt length matters less than clear logic and structure, but knowing how to frame a problem remains the most reliable way to bridge the gap between what you want and what the model hands back.

Miss out on any previous AI terms of the week? Check out our Glossary of AI terms, where we’re collecting all entries in the series. And if you want us to define a term in an upcoming newsletter, let us know!

The Innovation Profs are Drake University Professors Chris Porter and Chris Snider. Porter leads Drake’s Artificial Intelligence major and Snider is a multimedia professor in the School of Journalism and Mass Communication.