🤖 Shopify’s New Rule: Prove AI Can’t Do It First

Here are the top stories for today:

1. 💻 [FREE Webinar] From Misuse to Abuse - AI Risks and Attacks

2. 📸 Google’s AI Mode Just Got a Vision Upgrade—Literally

3. 🌟 Amazon’s Nova Reel Gets a Glow-Up

4. 🤯 Quick Takes

Read Time: 5 Minutes

From Misuse to Abuse - AI Risks and Attacks

Cybercriminals are not just leveraging AI but are also finding ways to exploit and bypass AI-based security measures.

Cato CTRL’s (Cyber Threats Research Lab) latest Cybersecurity Masterclass delves deep into how threat actors misuse and exploit AI technologies. Hosted by Etay Maor, Chief Security Strategist, and Vitaly Simonovich, Threat Intelligence Researcher at Cato Networks, this session provides unique insights into the methods cybercriminals use to target AI vulnerabilities.

What’s covered:

• Unveiling the reality beyond the headlines: How cybercriminals are truly using AI

• The exploitation of Generative AI by threat actors

• Real-world examples showcasing Generative AI in attacks and as a target

• The risks associated with private GPTs and their vulnerability to attacks

This masterclass equips cybersecurity professionals with essential knowledge to understand and combat the evolving threats posed by AI misuse. Stay informed, stay secure.

Google’s AI Mode Just Got a Vision Upgrade—Literally

Google’s AI Mode isn’t just flexing—it’s evolving fast. Initially rolled out to Google One AI Premium subscribers, the early buzz has been all about how clean, fast, and eerily intuitive it is. So, naturally, Google’s letting more folks in on the fun and throwing in a powerful visual twist.

What’s New in AI Mode:

• Multimodal Magic: Snap or upload a photo, ask a question about it, and get a full breakdown. Google’s Lens + Gemini tag team now powers this smart visual Q&A.

• Context is King: AI Mode doesn’t just identify objects—it gets the vibe. Think materials, colors, layout, relationships between things.

• Supercharged Search: Uses “query fan-out” to shoot off multiple related searches at once, giving you deeper, richer responses than standard Google.

• Smarter Recommendations: AI Mode can now ID each book on your shelf, then suggest similar reads with direct links to buy or learn more.

• Open Access via Labs: U.S. users can now test AI Mode in the Google app (iOS and Android).

TL;DR:

Google’s AI Mode is evolving from a clever chat tool into a full-blown visual search genius. Whether you’re asking complex questions, planning a weekend getaway, or wondering what the heck that mystery ingredient in your fridge is, AI Mode’s got you.

💡 Did You Know?

ChatGPT can now see, hear, and speak: OpenAI’s latest version of ChatGPT is multimodal, meaning it can understand images, respond with voice, and even hold natural-sounding conversations in real-time.

Amazon’s Nova Reel Gets a Glow-Up

Amazon’s AI video whiz, Nova Reel, just leveled up—and it’s not shy about showing off. The newly launched Nova Reel 1.1 can now stitch together up to two minutes of slick, AI-generated footage, with a consistent cinematic vibe that actually makes sense from shot to shot (a big win in AI video world). It’s also rolling out “Multishot Manual” mode, which lets you upload a reference image and a short prompt to fine-tune the scene. Think storyboard meets robot director. But while Nova’s video chops are growing fast, its transparency? Not so much. Amazon’s keeping quiet about what videos trained the model, which leaves some folks raising eyebrows (and possibly legal flags). Still, it’s available through AWS if you’re brave—or bold—enough to try it. Just don’t forget who’s behind the camera.

Quick Takes

• Shopify’s New Rule: Prove AI Can’t Do It First: In a company-wide memo, Shopify CEO Tobi Lütke introduced a striking new policy: before any team can ask for additional headcount or resources, they must first prove that AI can’t handle the task. This bold shift reframes AI not as a helper, but as a default teammate—one whose capabilities must be ruled out before bringing in human reinforcements. Lütke encouraged teams to imagine their departments with autonomous AI agents already embedded, sparking what he called “fun discussions and projects.” But beneath the playful tone lies a serious transformation: one that could significantly reduce Shopify’s need for human labor, especially as AI systems become more advanced.

• Meta’s Maverick Ranks High—But Not Without a Catch: Meta’s new AI model, Maverick, scored second on LM Arena, but it turns out the version tested wasn’t the same one available to developers. Researchers have flagged that Meta used a custom, chat-optimized variant for benchmarking raising eyebrows over transparency and real-world performance expectations.

That’s all for Today!

Advertise with Us

Want to reach an engaged audience of tech enthusiasts and AI professionals?

Our newsletter is read by thousands of dedicated subscribers who are passionate about the latest advancements in AI and technology. Contact us today to learn more.

What did you think of today’s newsletter?