New York Legislature Passed AI Disclosure Law Now Awaiting Governor Signature — Here's What That Means for the Rest of Us

NEW YORK'S AI DISCLOSURE LAW

NY legislature passed the first-in-nation bill requiring AI labels on news content

The Rundown: New York just made AI disclosure in news a legal requirement, not a best practice. Now awaits the Governor's signature.

The details:

• The NY FAIR News Act mandates clear disclaimers when published content is substantially or wholly AI-generated.
• The bill passed both chambers with bipartisan support and backing from SAG-AFTRA, Writers Guild East, and NewsGuild of New York.
• It now awaits Governor Kathy Hochul's signature.
• Sponsors framed the legislation as addressing declining public trust in media during ongoing attacks on the press.
• The law applies to news organizations operating in New York State.

Why it matters for us: If you have any New York readership — or if your state follows suit — you'll need labeling protocols for AI-assisted content. But flip this: it's also a marketing angle. 'Written by humans who live here' becomes a competitive differentiator you can actually prove.

Read at nysenate.gov →

WHO'S LIABLE WHEN AI LIES

Germany made Google liable for its AI's lies — could the U.S. be next?

The Rundown: A Munich court just ruled that when Google's AI Overviews invent false claims, those are Google's own words — and Google can be sued for them. The open question this raises for America: whether the same logic could finally pierce Section 230, the legal shield that's protected U.S. tech platforms for 30 years.

The details:

• A Munich court classified Google as a "direct infringer" after its AI Overviews falsely tied two publishers to scams and subscription traps.
• The court's reasoning: AI Overviews produce "independent, new, and substantive statements" in Google's own words — making them Google's speech, not a neutral list of links.
• The false accusations appeared nowhere in the real search results — the AI invented them, then presented them as fact.
• Google must cover 80% of the legal costs and is barred from repeating the false claims; it confirmed it will appeal on June 12, 2026.
• The U.S. difference is Section 230, which since 1996 has shielded platforms from third-party content — but AI Overviews are generated by the platform, possibly outside that shield.

Why it matters for us: This cuts both ways for a community paper. If Google's AI Overview ever invents something false about your publication or garbles your reporting, this is the first ruling anywhere that points to a path for fighting back. But it's also a warning: if you run AI-generated copy that gets a local business or person wrong, a court may decide those are your words — not the AI's.

Read at wired.com →

AI SCAM PIPELINE

Scammers are planting fake numbers for Google's AI to serve your readers

The Rundown: That helpful phone number in Google's AI summary? Might connect grandma to a fraudster.

The details:

• Scammers publish fake customer service numbers across low-profile websites, which Google's AI then scrapes and presents as legitimate.
• Victims searching for company contacts get connected to fraudsters who steal payment and personal information.
• Credit unions and banks are actively warning customers about these AI-enabled scams.
• Google's advice: perform additional searches to verify. There's no way to disable AI Overviews.
• The AI interface presents scraped information as authoritative fact, increasing victim susceptibility.

Why it matters for us: This is a ready-made consumer protection story for your readers — especially seniors. Write a local explainer this week with the specific advice: always verify phone numbers on official company websites, never trust the number an AI summary hands you.

Read at wired.com →

CHATGPT'S SUPERAPP PIVOT

OpenAI redesigning ChatGPT into an everything-app with embedded services

The Rundown: ChatGPT is becoming less chatbot, more app mall with a typing box.

The details:

• OpenAI plans to integrate coding tools, AI agents, image generation, and partner services like Canva and Booking.com directly into ChatGPT.
• Business customers now represent about 40% of OpenAI revenue, expected to hit 50% by year-end.
• Roughly 2 million businesses currently use ChatGPT.
• Changes could appear on ChatGPT's website and mobile apps within weeks.
• OpenAI has been linked to a potential confidential US IPO filing.

Why it matters for us: When readers can book travel, design graphics, and finish tasks without ever opening a browser, 'being discoverable on the web' matters less. Start asking now: how does your paper show up inside these AI platforms — or does it just disappear?

Read at mindstream.news →

What is?

Agentic

What it is: The buzzword of the year, and it just means AI that does things instead of only talking about them. A regular chatbot is reactive — it waits for your next question. "Agentic" AI is proactive: hand it a goal and it takes the steps to reach it on its own, using tools, making decisions, and working until the job is done. When you hear "agentic AI" or "the agentic era," that's the shift being described — from AI you chat with to AI that goes and does the work.

Why publishers care: It's the single word sitting underneath almost every big AI headline right now, including ChatGPT's everything-app pivot in this very issue — so knowing it lets you read AI coverage fluently instead of nodding along. More practically, it's a filter: agentic tools are the ones that actually save a newsroom hours, because they don't just draft a thing, they execute it — pull the data, build the file, schedule the post. When you're deciding which new AI tool is worth your time, "is it actually agentic?" is the question that separates real time-savers from another chatbot.