This week, we’ve got 5 under the radar stories that shaped AI this week, from a new video model and a free assistant browser to culture trends and a courtroom move with big ripple effects. Then we zoom into the Tilly Norwood debate about synthetic talent. Let’s dive in ⬇️

In today’s newsletter ↓
🌟 5 stories that deserve a second look
🎭 A synthetic star tests industry boundaries
💸 OpenAI hits $5 billion valuation
💎 Google unlocks CLI and API for Jules
🎯 Weekly Challenge: Pressure test viral news

Here are five stories that mattered more than their headlines suggested. From a free assistant browser and rising AI romance to a Texas classroom experiment and a courtroom fight, these signals show where everyday AI is actually going. Sora 2 is in the mix, but not the lead.

A new U.S. survey made waves: roughly 28% of adults say they’ve had an intimate or romantic relationship with an AI chatbot, while a majority report some relationship category with AI (from “friend” to “colleague”). 

That’s a cultural step change, raising questions about consent, disclosure, and the emotional design of agents. Even if sampling and definitions deserve scrutiny, the directional signal is clear: relational use of AI is moving into the mainstream, with implications for safety policies and product UX. 

OpenAI unveiled Sora 2, a flagship video-and-audio model with synchronized dialogue and effects, better physical plausibility, and improved multi-shot control. There’s also a new social Sora app for short clips, remixing, and verified self-insertion “cameos,” with an API on the roadmap. 

The big bet: model “world physics” well enough that failure looks like failure, not a glitch. If Sora 2’s realism holds up beyond demos, creative pros will watch for licensing terms, watermarking, and longer durations. The near-term impact is on short social formats; the medium-term test is whether it slots safely into professional pipelines.

Perplexity made its Comet AI browser free for everyone, moving from a $200 Max-only perk to an open, rate-limited tier. The company also introduced Comet Plus, a $5 add-on (included with Pro and Max) that surfaces curated reporting with revenue sharing for publishers. 

The goal is to counter low-quality “AI slop” and make assistant-first browsing mainstream. Early coverage frames this as a direct challenge to Chrome, and Perplexity says the free tier is “indefinite.” 

If Comet’s sidecar assistant really reduces tab-juggling for research, shopping, and trips, it could nudge casual users toward an always-on copilot model.

Alpha School’s AI-first model returned to headlines as outlets profiled classrooms where learners spend about two hours on core academics using AI-driven software, while high-paid adult “guides” coach motivation and life skills. 

Supporters tout faster mastery and student engagement; critics worry about equity, quality control, and what’s lost when teachers are reframed as facilitators. Recent features show the approach expanding beyond Texas, keeping the “teachers vs AI” debate very current in K-12.

In the continuing rivalry with Elon Musk’s xAI, OpenAI asked a federal judge to dismiss claims that it poached employees to steal trade secrets, arguing the suit is baseless and that employees may change jobs. 

It’s a reminder that AI competition is playing out in courtrooms as much as in benchmarks and product updates. Watch for how the judge treats the trade-secrets narrative and mobility rights, which could shape recruiting norms across the sector. 

A photoreal synthetic performer named Tilly Norwood drew industry wide backlash after her creators pitched her as a bankable star. The project comes from Xicoia, an AI talent studio tied to Particle6, and was promoted around a short comedic sketch and a provocative promise to scale synthetic performers.

Reports say talent agencies are exploring representation even as unions and A list actors condemn the idea. Read background on the AI actress push and the early talent talks. 

Guilds and stars warned that a synthetic actor risks erasing lived experience, consent, and fair pay. UK union Equity called Tilly an AI tool rather than a performer, while coverage tallied reactions from high profile names who see a slippery slope toward replacement. 

Commentaries in major outlets argued the figure embodies a broader unease with authenticity in film culture. Scan the union response and an industry wide critique.

Key questions now shift to provenance, training data, and likeness consent. Creative leaders will look for clear crediting, pay, and opt in controls before any studio scale use. International reaction suggests a patchwork of standards unless guilds and studios align with regulators. For a global snapshot, see coverage from European outlets rounding up the flare up.

🦠 AI helps map how antibiotics target gut bacteria

🧬 HHS doubles funding for AI in childhood cancer research

💸 OpenAI hits $500B valuation after employee stock sale

🤝 Fujitsu and NVIDIA broaden their collaboration

📈 AI startup valuations spark bubble fears as funding increases

🧑‍💻 Google’s AI coding agent Jules adds CLI and API

It’s becoming increasingly hard to know what’s real and what’s not, especially if your LLM tells you something with confidence. This week, we’re equipping you with the tools to quickly stress test viral claims, so you can be more confident about the information you’re learning.

📃 How to do it:

That’s it for this week! Did we miss any major AI stories that you want to learn more about? And, what are your thoughts on Tilly Norwood? Hit reply and let us know your thoughts.

Zoe from Overclocked