Hi Friend,

The conversations that I seem to keep having are around lead/deal flow/sales, and AI SEO/GEO. We’ll be posting a recording about AI SEO/GEO after the live session for the Global AI Council. Look for it on our YouTube or in the newsletter next week!

This week’s AI theme is like last week: real deployments and real guardrails. We are seeing AI roll into hospitals and factories, while regulators and courts keep tightening the “rules of the road” for how these systems get built and used.

Quick question for you as you get going in Q1: where would a little more speed help most right now? Is it customer support, operations, compliance, or content creation? Keep that answer in mind as you skim the stories below.

OpenAI brings ChatGPT into hospitals with a healthcare-focused rollout

OpenAI announced “ChatGPT for Healthcare,” aimed at helping clinicians with clinical and administrative work. The rollout hits major health systems including AdventHealth, HCA Healthcare, Cedars-Sinai, and Stanford Medicine Children’s Health. The approach highlights HIPAA-oriented controls, including options like customer-managed encryption keys and citations to clinical sources to support fast verification.

My take: Healthcare is one of the toughest environments for AI because the bar for accuracy and privacy is so high. For any business selling into healthcare, this is a strong signal that “enterprise-ready” now means workflow fit, auditing, and security design, not just model quality. We have created a HIPAA compliant AI for women where it’s trained only on research data from women’s studies. You can test it out here at CycleWise.ai and get on the wait list for the version with your own account. It will be free to uterus owners forever!

Read more here

MIT’s practical test for whether clinical AI is memorizing patient data

MIT researchers shared a testing setup to evaluate whether models trained on de-identified health records can still “memorize” patient-specific details and leak them under targeted prompting. The key idea is to measure privacy risk in context, including how much outside knowledge an attacker would need for a leak to be harmful.

My take: This is the kind of practical research that moves AI from “experimental” to “deployable.” It helps organizations buy with confidence by addressing the specific fear of data memorization and leakage. Which is a huge fear right now, that we often hear about. If you are procuring tools in sensitive fields, you should be asking vendors for this level of “leakage testing” alongside your standard security checks to ensure your data stays yours.

Read more here

Siemens + NVIDIA push industrial AI with digital twins, copilots, and a PepsiCo case study

Siemens and NVIDIA expanded their partnership around an “Industrial AI Operating System,” with Siemens also introducing Digital Twin Composer and a set of industrial copilots across the value chain. Siemens highlighted PepsiCo using the digital twin approach to simulate facility upgrades and optimize throughput before making physical changes.

My take: The real business story here isn’t the technology itself, but the ability to avoid expensive mistakes. By using digital twins to simulate changes, operations teams can turn high-stakes decisions into safe experiments. It’s a reminder that the best ROI often comes from using AI to predict and prevent issues in the digital world before they cost you money in the physical one.

Some of my friends have created a virtual board where they can run ideas by a diverse set of experts. You can do this one too!

Read more here

US AI regulation: new state rules are live, and compliance is getting more specific

A wave of US state tech laws took effect at the start of 2026, including multiple AI-related requirements such as transparency obligations in California and additional state activity around elections, chatbots, and disclosures. The Verge also notes a court action that paused one high-profile Texas age verification rule, which is a reminder that some of these laws will be tested quickly.

My take: If you run marketing, product, or legal, this is your cue to treat AI like any other operational risk: document what tools you use, what data they touch, and what you disclose to customers. Even a lightweight “AI inventory” can save you from last-minute scrambles when a client asks, “Are you compliant?”

You can keep it super simple with a list in a doc of all the tools that you use and keep a folder of the training data if you are creating custom gpts and agents. This will make it easy to give your “evidence” if you are audited or if a client wants to know. Bonus point for keeping links to the use policies of each tool in that list too. Your customer may now want to know if you know how safe their data is in your tech stack.

Read more here

US courts and AI: copyright fair use questions are heading into a pivotal year

Reuters reports that US courts are weighing whether training AI on copyrighted content qualifies as fair use, with different judges signaling different approaches and major lawsuits still moving through the system. The uncertainty is already pushing some companies toward licensing deals and negotiated settlements, even while more decisions are expected in 2026.

My take: This matters even if you are not building models. If your team uses generative AI for marketing, creative, or product content, your safest move is to be intentional about sourcing, licensing, and provenance, and to choose vendors who can explain their data posture without hand-waving.

Your responsiblility will be in identifying your own agent/customgpt training data. A liability could be as simple as someboy on your team uploading a book as training data that they don’t have the legal right to upload.

Read more here

Tools to try this week

• Pictory.ai –  Create and edit professional quality videos (for social media, blogs, and other marketing materials) using text in minutes.
• Open Forge AI – AIO platform for mastering AI search, leveraging advanced AISEO and GEO-targeted insights to ensure your brand dominates the competition and wins more customers.
• Suno AI – AI-powered music creation platform that empowers anyone to instantly generate high-quality, and personalized songs from simple text prompts.
  

Applied AI – Level 1 Certification

We’ll start with the basics to build a solid AI foundation before diving into advanced topics.

Module 1: Introduction to AI

Learn essential tools, terms, and how to fact-check AI content. You’ll also create a simple AI use policy to ensure responsible usage.

Module 2: Prompt Engineering and Brand & Personal Voice

Master crafting effective prompts and maintaining a consistent brand voice across AI tools, so everything you create stays on-brand.

Don’t just take our word for it…

Here’s what a recent student had to say:

“I highly recommend this course to anyone in who’s been curious about how to use AI in a practical way. The instructors, Nicole, Andrea, and Maricar, brought a well-structured approach to every session. I had been looking for guidance on which tools to use and how to use them effectively and they delivered that as well as many ideas and tips. If you’re navigating AI and want to understand what tools will support your work and how to use them, this course is absolutely worth it!”

– Charis Armer, Business Consultant

Ready to get started?! Our next 10-week live course begins on last week of January

Get on the Waitlist Today!

How to Turn Images into Video with Pictory

Do you have a folder full of images but need a compelling video? Pictory allows you to skip the complex editing process by instantly converting a group of static images into a polished video timeline.

Follow this guide to create your video in minutes.

1. Import and Scene Creation

The process begins with your raw assets.

• Upload your images: Select the group of images you wish to use.
• Automatic Scene Generation: Pictory will instantly turn each image into a distinct video scene. You don’t need to manually place them on a timeline; the software handles the initial sequencing for you.

2. Add Your Narrative

Once your scenes are generated, you can weave in your story.

• Input your Script: If you have a prepared script, you can simply copy and paste sentences directly into the storyboard area.
• Caption Styling: The text from your script will appear as captions. You can fully customize these by adjusting the fonts and colors to match your brand style.

3. Customize Your Visual Layout

This is where you make the video unique. You aren’t stuck with a static slide show; you can shape the “look and feel” of every frame.

• Create Custom Layouts: You can manipulate the scene layout to suit the image.
• Move Images: Reposition your photos within the frame for better focus.
• Add Decor & Transitions: Enhance the visuals by adding “decor text” (visual text elements) and applying transitions to your text to keep the viewer engaged.

4. Set the Tone with Audio

A video isn’t complete without sound. Pictory allows you to layer audio easily:

• AI Voiceover: Add an AI-generated voice to narrate your script automatically.
• Background Music: Select and add a backing track to establish the emotional tone of the video.

Pro Tip: If you have captions or text embedded in your original images that clash with your new subtitles, you can add elements to scenes to “block out” or cover up the underlying captions.

5. Preview and Download

Before you finish, ensure everything flows correctly.

• Preview: Watch the video within the editor to check the timing, voiceover sync, and visual transitions.
• Download: Once you are satisfied, click Download. Pictory will render a finished, shareable video file for you.

Join the Conversation

Have questions, comments, or cool AI stories to share? We’re all ears and always looking to highlight insights from the AI community. Simply reply to this email or drop us a line at info@aismartventures.com.

Can’t wait to hear from you!

Cheers,

-Nicole Donnelly
Founder, AI Smart Ventures