🤓 What Happens When AI Joins Your Meetings

PLUS: This Week in AI

We thought AI was here to take notes.

Instead, it’s taking the space where real talk used to happen.

Meetings have always been a struggle—too long, too unfocused, too often calendar fillers with no real takeaways.

Now, AI notetakers are promising to fix that.

More companies are bringing AI into meetings to save time, improve follow-ups, and make things more inclusive.

Microsoft has one. Google has one. Zoom, Otter, and dozens of startups do too.

They transcribe every word.

Summarize the discussion.

Highlight key decisions.

Draft action items.

In theory, it’s a win: fewer distractions, fewer “wait—what did we decide?” moments, and more structured takeaways.

The Upside (on paper):

✅ Stay engaged without scrambling to take notes
✅ Let people who missed the meeting catch up
✅ Auto-generated action items and summaries
✅ Searchable records of what was said

It’s efficient.

It’s organized.

And it feels like progress.

Ok, so what’s the catch?

The Tradeoffs (No One’s Talking About)

⚠️ It changes how people behave.
Once you're being recorded, your guard goes up. You self-edit. You sand down the sharp edges.

The result:

• Risky ideas vanish

• Debate softens

• Real talk moves to Slack (or the parking lot)

⚠️⚠️ The AI might capture your words—but miss your meaning.
Sarcasm, hesitation, nuance, silence—all the things that shape a conversation don’t always make it into the summary.

⚠️⚠️⚠️ There’s no real consent.
A pop-up saying “this meeting is being recorded” isn’t a real choice. And “you can opt out” isn’t real if no one feels like they actually can...

⚠️⚠️⚠️⚠️ The data goes... where, exactly?
Who owns it? Who can access it later—and for what use cases?
Can it be used to train AI models? Can your company use it to review performance? Can it be subpoenaed?

And who else might see it—beyond the people in the room?

Most companies don’t say. Most employees don’t ask.

Image courtesy of: ChatGPT & Avi

Just to be clear—I’m a big fan of AI note-takers. But using them well takes more intention than most people realize.

If You’re Going to Use an AI Note-Taker, Do It Right

✔️ Ask for consent before the meeting, not while it’s starting

✔️ Make “no” a real option.
If people can’t opt out without consequences or feeling awkward, it’s not a choice—it’s pressure. Have a backup: someone taking notes.

✔️ Be clear about what’s being captured.
Is it just voice? Or video? Does it pick up body language and facial expressions too? If it’s recording more than people expect, be upfront about it.

✔️ Share where the data lives, and how it will be used.
Who beyond the people in the room will see the notes? People calibrate what they say based on who’s in the room. If the audience isn’t clear, honesty takes a backseat to caution.

Clarity builds trust. When the policy is clear—that notes are just for those who couldn’t attend, won’t be shared elsewhere, and will be deleted within a week—it changes how people show up.

✔️ Scan for missed context.
AI-generated summaries can miss tone, emotional nuance or subtext—especially during tense, sensitive or high stakes conversations.

—

We wanted AI to capture what was said.

Instead, it’s changing what we say—and how we say it.

The most valuable meetings aren’t the ones with perfect information exchange.

They’re the ones where someone finally says the quiet part out loud.

---

---

What You Need to Know About AI This Week ⚡

Clickable links appear underlined in emails and in orange in the Substack app.

---

🎶 Universal Music Group, Warner Music Group, and Sony Music Entertainment are negotiating licensing deals with AI music startups Udio and Suno.

Yes—the same startups they’re suing for copyright infringement.

Instead of waiting for a courtroom verdict, they’re eyeing settlements that include licensing fees AND equity.

Why the shift?

Because the tech (and consumer behavior) is moving faster than the courts.

And there’s too much money on the line.

A drawn-out legal battle helps no one.

Labels want compensation and control. And to avoid another Napster moment.

AI startups want flexibility—and survival. And to avoid becoming the next Napster.

This seems to be the emerging pattern:

First you sue. Then you settle. Then it becomes the playbook.

---

🆕 ChatGPT Adds Meeting Recorder + Smart File Access

(But availability varies across plans)

🙌 Built-in Meeting Recorder and Transcriber: Captures meetings, brainstorming sessions, and voice notes—up to 120 minutes per session—automatically generating transcripts, summaries with time-stamped citations and suggested action items. It also lets your query transcripts so you can use them as a resource to create other documents.

Recordings and transcripts are not used for model training, and audio files are deleted post-transcription.

🙌🙌 Search & Deep Research Connectors: Lets you pull info and real-time context from your own internal data, documents, or third-party tools and services. Plus, it’ll show exactly where each piece of information came from, so you can quickly double-check it.

Supported Connectors: Google Drive, Outlook, Teams, Gmail, Google Calendar, Box, Dropbox, SharePoint, Linear, GitHub, and HubSpot

⚠️ Availability & Limitations

• Meeting Recorder: Currently exclusive to Team users on macOS, but expanding soon to Plus, Pro, Enterprise, and Edu users in the near future.

• Connectors: Availability varies; some connectors like Dropbox, SharePoint and HubSpot are limited to Team, Enterprise, and Edu users. You can find the details here.

• Regional Restrictions: Certain features are not available in the EEA, UK, and Switzerland.

---

Reddit is suing Claude maker Anthropic for scraping its site to train AI models without payment or consent. It says it tried to reach a licensing deal but couldn’t.

Reddit already has lucrative licensing deals with OpenAI and Google.

---

While most AI companies are cozying up to Trump’s White House, Claude maker Anthropic is lobbying against its key AI bill—one that would block states from setting their own rules.

The move has frustrated administration officials, who also took issue with CEO Dario Amodei’s recent prediction that AI could wipe out half of entry-level white-collar jobs within five years.

But in staying out of step, Anthropic may be setting itself up for the long game—reinforcing its appeal to researchers, recruits, and the next administration.

---

🍿 Every now and then, a story comes along that feels too wild even for Hollywood.

When Sam Altman got ousted from OpenAI, I called every producer friend:

“This one has everything—secret alliances, back-room betrayals, power plays, and egos the size of small planets. And at the heart of it stands the man who’s quietly becoming one of the most powerful people in the world.”

Turns out I wasn’t the only one.

Luca Guadagnino is set to direct Artificial for Amazon MGM, based on that five-day shakeup when Altman was fired—and then reinstated after a full-on revolt.

Word is Andrew Garfield’s circling Altman, Monica Barbaro’s eyeing badass CTO Mira Murati, and Yura Borisov could play Ilya Sutskever—the co-founder who led the coup and the comeback.

I was deep into every nuance of this saga in real time.

The headlines barely captured a fraction of the unseen fireworks. If this gets even a part of the untold drama right, it can out-Social Network The Social Network.

P.S. Amazon/MGM, if you need an AI obsessed consultant who also knows how to market movies, hit me up. I know this audience like the back of my hand. 😃

---

Google’s new video model, Veo 3, can generate convincingly fake videos of riots, election fraud, racial violence—with minimal prompting. Experts warn it could fuel chaos during breaking events—especially since real clips are already being mistaken for AI.

---

Unemployment is rising among recent grads—especially in tech and finance—as more companies adopt an “AI-first” approach to hiring. Many now skip junior roles altogether, betting AI can do the work.

This shift has led to fewer training and mentorship opportunities, with some grads skipping traditional paths entirely to pursue risker paths like starting their own companies or joining startups.

---

Everyone in Hollywood is using AI. They’re just hiding it.

This piece from Vulture explores how generative tools are already being used—to write pitches, generate mood boards, cut trailers, and patch budget gaps—all behind closed doors.

It lines up with what I see daily: many execs I advise ask that our conversations stay off the record.

🌟 Some highlights:

• Artists are being asked to “clean up” AI-generated concept art—so it doesn’t look like it came from AI.

• Studios are generating trailers for films they haven’t even shot to secure financing.

• Executives are asking AI to repurpose existing IP into new formats—like anime versions, PG-13 cuts, or entirely new formats.

• Directors are using AI to storyboard multiple versions of a shot—on deadline, without a team.

• Visual effects teams are using AI to simulate explosions, landscapes, and effects, at a fraction of the cost.

Because so much of this happening behind closed doors, few people—inside or outside the industry—really grasp how fast things are moving.

Driving this transformation in the shadows ensures we’ll all be blindsided by their consequences.

---

AMC just became the first cable network to sign a formal deal with AI video firm Runway. It’ll use the tech for pre-viz and consumer-facing marketing materials—but not in shows (yet).

Given how fast these tools are evolving, I keep wondering what these deals are actually buying—and why they’re even necessary.

I know the Lionsgate deal was a “no money” one. If this is just about making sure Runway doesn’t train its main models on the studio’s IP, couldn’t that be handled in a standard service agreement—especially since the deals aren’t even exclusive?

So, perhaps it’s just about the headlines? Anyone?

---

😅 People are asking ChatGPT if they’re hot enough.

Users are uploading photos and asking for brutally honest feedback on their looks—then spending money to fix what it flags.

They say it’s more “objective” than friends.

But the advice is shaped by biased training data, commercial incentives, and a narrow view of beauty—often echoing the internet’s loudest (and harshest) voices.

In case you missed last week’s edition, you can find it 👇:

That's all for this week.

I’ll see you next Friday. Thoughts, feedback and questions are always welcome and much appreciated. Shoot me a note at avi@joinsavvyavi.com.

Stay curious,

Avi

💙💙💙 P.S. A huge thank you to my paid subscribers and those of you who share this newsletter with curious friends and coworkers. It takes me about 20+ hours each week to research, curate, simplify the complex, and write this newsletter. So, your support means the world to me, as it helps me make this process sustainable (almost 😄).