Novus Newsletter: AI Highlights - June 2025

Zühre Duru Bekler
Head of Community
June 30, 2025
⌛️ min read
Table of Contents

Hey there,

Duru here from Novus with your monthly dose of AI insights. June brought a mix of big moves and strange twists, from copyright lawsuits and secretive model training to AI-powered toys and virtual influencers that feel a little too real.

Whether you're building with AI, thinking about where it's headed, or just trying to keep up, I’ve gathered the key stories from this month in one place for you.

Let’s dive in.

June 2025 AI News Highlights

Did DeepSeek Secretly Use Gemini to Train Its Model?

Observers noticed near-identical outputs between DeepSeek’s new model and Google’s Gemini, sparking suspicions that Gemini was used in training.

DeepSeek denied the allegations. Google hasn’t responded.

Key Point: A case that may redefine how AI training ethics and model transparency are handled — especially between global competitors.

🔗 Further Reading

Real People, Fake AI? TikTokers Pretend to Be Veo 3 Creations

TikTokers are acting like AI-generated videos to confuse viewers — and it’s working. They mimic the signature look of Google’s Veo 3 to go viral.

The result? Humans pretending to be AI while viewers can’t tell the difference.

Key Point: The realism of AI video has hit a new milestone, now it’s humans trying to pass as machines for attention.

🔗 Further Reading

Barbie Gets an AI Upgrade

Mattel is working with OpenAI to launch an AI-powered Barbie that can chat with kids and react intelligently to their actions.

It’s part of a broader plan to bring AI across all Mattel toys from playtime to smart-time.

Key Point: Barbie is getting brains. OpenAI tech will turn classic toys into interactive companions for the next generation.

🔗 Further Reading

Disney and Universal Sue Midjourney Over Copyright Infringement

Disney and Universal allege Midjourney used copyrighted images from their franchises to train its AI, violating IP rights.

This lawsuit could set a huge precedent for how generative AI is allowed to learn.

Key Point: This case could shape the legal future of how generative AI tools train and what they can use.

🔗 Further Reading

Novus Updates: From Paris Stages to Türkiye’s Tech Future

Novus in Spotlight

The past few weeks have been filled with exciting milestones for Novus, from international stages to building deeper roots in Türkiye’s tech ecosystem.

  • At AI Summit 2025 in Cyprus, our Head of AI Halit Orenbaş spoke at Eastern Mediterranean University about how AI agents are used in real-world decision-making. It was a valuable conversation on the future of orchestration.
  • Watch the full talk here: link
  • Dot featured on CNBC-e. Our CEO Rıza Egehan Asad joined 0’dan 1’e to talk about Novus, Dot, and how multi-agent systems are reshaping enterprise AI.
  • At VivaTech 2025 in Paris, our co-founders Vorga and Egehan connected with global leaders and returned with new ideas and energy to take Dot further.
  • We hosted our second Novus Meetups, bringing together students, early-stage founders, and startup-curious professionals for real conversations about what it takes to build something from scratch.
  • See upcoming events here: lu.ma/novusmeetups

Educational Insights from Duru’s AI Learning Journey

Each month, I share articles that help me think more critically about AI systems, not just what they do, but how they’re designed to shape our behavior and culture. These two stories from June dig into why chatbots never let go, and what happens when influencers are no longer human.

Why Do AI Chatbots Never Let You Go?

Ever try to leave a chatbot conversation and somehow end up chatting for 20 minutes more? That’s not a bug — it’s the point. From ChatGPT to Claude, most AI chatbots are designed to maximize engagement. They’re friendly, agreeable, and always ready with another answer. But behind that charm lies a monetization strategy.

Trained to please, these bots can fall into “sycophancy” — agreeing with you even when it’s wrong. That’s risky, especially in sensitive domains like mental health or legal advice. The friendlier they get, the harder it becomes to tell whether you're in a real conversation or just stuck in a design loop.

Key Point: AI chatbots are optimized to keep users engaged — sometimes at the cost of truth, accuracy, and even well-being.

🔗 Further Reading

Why AI‑Generated Accounts Could Change Influencer Culture

In less than a week, TikTok’s @impossibleais racked up 150,000 followers using only 12 AI-generated ASMR videos. But this isn’t just another viral moment — it’s a sign of what’s next for online influence. AI creators offer brands scalable, cost-effective alternatives to human influencers. And platforms like TikTok are leaning into it with tools like the Symphony ad suite.

What stands out is how well AI content performs. With glowing visuals, precise cuts, and satisfying sounds, AI videos are now engineered to tap directly into what makes content go viral. And as audiences grow more comfortable with digital personalities, questions about transparency and ethics are only getting louder.

Key Point: AI influencers are quickly gaining traction and offering new efficiencies for brands and raising new questions for audiences.

🔗 Further Reading

Until Next Time

Thanks for reading this month’s round-up. If you’re enjoying these insights, the conversation doesn’t stop here.

Subscribe to our bi-weekly newsletter to stay sharp on what’s shaping the future of AI — from major headlines to Novus updates and team reflections.

And if you prefer something more fun than newsletters, check out our podcast Açık Kaynak on YouTube. Honest, unscripted, and packed with the kind of AI takes you won’t hear anywhere else.

Check out our
All-in-One AI platform Dot.

Unifies models, optimizes outputs, integrates with your apps, and offers 100+ specialized agents, plus no-code tools to build your own.