Host Jason Michael Perry sits down with Justin Pierce, co-chair of the Intellectual Property Division at Venable LLP, to talk about who actually owns your ideas once AI is involved, and why the answer changes depending on where you typed them.

Anthropic just paid over a billion dollars to settle the largest copyright case in U.S. history, over digital books used to train Claude. The twist: the same judge ruled that scanning physical books the company had legally purchased counted as fair use. Jason and Justin unpack where that line actually sits, why a dozen similar lawsuits against AI companies are still unresolved, and what it means when the companies fighting over stolen training data built their own models the same way.

They also get into who owns the knowledge in your head once you write it down for an AI to use, why your ChatGPT history might be more discoverable than your email, and what Justin tells CEOs about what never to type into a chatbot.

Podcast Notes & Links

• Anthropic settlement + the Alsup fair use ruling
• Apple v. OpenAI (trade secrets)
• The New York Times v. OpenAI
• Meta's Model Capability Initiative
• Google v. SerpApi dismissal
• Learn more about Venable LLP
• PerryLabs - Start with an AI Assessment

Sponsors

This episode is brought to you by:

PerryLabs: Invested in AI but not seeing results? You're not alone. PerryLabs helps organizations find what's working and build what's next. Start with our AI Assessment at PerryLabs.io.

HoopDee
Are you a new parent, or do you have one on the way? Get HoopDee so you never waste a drop of breast milk or baby formula. HoopDee's smart system does all the milk math for you, and the whole family will know which bottle to use next.

Enjoy 20% off with code: thoughts.
Visit gethoopdee.com.

Subscribe & Contact

Subscribe to the Thoughts on Tech & Things newsletter: jasonmperry.com/newsletter
Send feedback, questions, or guest suggestions: jasonmperry.com/contact or leave a message in the WYPR app or call 410-735-1616.

Credits

Thanks to the team at WYPR, our producer Sam Bermas-Dawes, and Myrna Martinez, Head of Operations and Marketing at PerryLabs.

Host Jason Michael Perry sits down with Greg Morton, data editor at The Baltimore Banner, to talk about what AI is actually doing inside a newsroom — and why admitting you used it can cost you a reader.

Long before anybody called it AI, machines were writing sports recaps, earnings reports, and election results. What changed isn't that machines write, it's that now they can write about anything and you can't tell. Greg and Jason get into the Banner's published AI rules and what they mean on deadline, what happens to what people believe when the good reporting is behind a wall and the junk is wide open, and how newsrooms are using these tools to get through document dumps no human could read in a lifetime. Greg was part of the team named a 2026 Pulitzer finalist for an investigation that tracked every city bus in Baltimore, every five seconds, for four months, to show why thousands of kids can't get to school on time.

Podcast Notes & Links

• The Baltimore Banner
• Greg Morton's work – The Baltimore Banner
• Missing the Bus: the Banner's school transit investigation
• 2026 Pulitzer Prize finalist citation, Local Reporting
• Ars Technica retracts an article containing AI-fabricated quotes – Nieman Lab
• How newsrooms used AI to work through the Epstein files – Reuters Institute
• Newswriter guilds on The Baltimore Sun's AI-written analyses – WYPR Midday
• BannerX 2026 – September 30 in Columbia, Maryland

Subscribe & Contact

Subscribe to the Thoughts on Tech & Things newsletter: jasonmperry.com/newsletter
Send feedback, questions, or guest suggestions: jasonmperry.com/contact
Leave a message in the WYPR app or call 410-735-1616

Credits

Thanks to the team at WYPR, our producer Sam Bermas-Dawes, and Myrna Martinez, Head of Operations and Marketing at PerryLabs.

Host Jason Michael Perry sits down with Chris Brandt, CEO of UpSurge Baltimore — the innovation arm of the Greater Baltimore Committee — and Managing Director at Audacious Capital, to unpack a growing fight over who controls the intelligence businesses increasingly rely on.

They discuss the tradeoffs between open and closed AI models, from guardrails and censorship to security, trust, and control. The conversation also explores AI distillation, the economics of building versus buying, and why companies may need to think beyond token costs when deciding how much of their intelligence they want to own.

They close by bringing the global debate over AI sovereignty home to Maryland and what it could mean for the region's businesses, investors, and entrepreneurs.

Podcast Notes & Links:
- Who Owns Your Company's Intelligence? — Chris Brandt, UpSurge Baltimore newsletter
- WTCI AGILE Global Innovation Series, Sept 23
- Meta introduces Muse Glimmer
- Anthropic to watermark Claude output under EU rules — The Register
- Anthropic restores Fable 5 with new jailbreak filter — The Hacker News
- Censorship in Chinese AI models can be undone — Semafor
- Kimi vs DeepSeek censorship comparison — ellamind
- OpenAI's gpt-oss open-weight models (Aug 2025) — MIT Technology Review
- PerryLabs - Start with an AI Assessment

Sponsors
This episode is brought to you by:

PerryLabs: Invested in AI but not seeing results? You're not alone. PerryLabs helps organizations find what's working and build what's next. Start with our AI Assessment at PerryLabs.io.

WTCI
The path to growth shouldn't stop at the border. The World Trade Center Institute provides the connections, knowledge, and global network Maryland's most ambitious leaders need to scale. From executive development to international trade, we help our region's businesses thrive on the world stage. Start your global journey at wtci.org.