iiinova — where your digital self keeps evolving

Full episode: youtu.be/yeOMuytoqgk?...

“I can say all the nice things about myself that I want.” Heather Physioc, Chief Discoverability Officer at VML, on our podcast about why credibility needs someone else to back it, and the marketing accounts running on fake followers.

100

Daypart.ai @daypartai.bsky.social · 09/10/2026

Full episode: youtu.be/yeOMuytoqgk?...

000

Daypart.ai @daypartai.bsky.social · 09/10/2026

Browser security firm Island found fake AI ad tools wearing ChatGPT and Claude branding, built to steal media buyers’ ad account logins. VML’s Heather Physioc: “It just creates this illusion of rigor even when the evidence is actually weak.”

110

Daypart.ai @daypartai.bsky.social · 07/10/2026

Our latest episode has Heather Physioc, who runs discoverability at VML, on how junk sites keep showing up next to real stores. Vibe coded site, fake Reddit consensus, official look. "People are so sick of getting duped by AI."

100

Daypart.ai @daypartai.bsky.social · 21/09/2026

Full episode: youtu.be/B7dtayeJQ-I?...

000

Daypart.ai @daypartai.bsky.social · 21/09/2026

In this clip, Shoshana Wodinsky shares how one sports betting site tried to stop users from leaving. When a cursor moved toward the corner of the screen, the site surfaced upcoming games based on that person’s betting history. If you bet on Brazil, you saw Brazil matches.

100

Daypart.ai @daypartai.bsky.social · 17/09/2026

Google can write the product claim and send the ad into market. The advertiser is still the one answering for whatever it says. Ad fraud can start before the click if the ad itself is making things up.

000

Daypart.ai @daypartai.bsky.social · 16/09/2026

Our podcast guest, privacy reporter Shoshana Wodinsky, followed a pregnancy audience back to phone-location data near baby stores. The seller said it covered 80,000 devices, with no proof about the people carrying them. Full episode in the link below.

100

Daypart.ai @daypartai.bsky.social · 15/09/2026

SB 690's bill history has it presented to California's governor Sept 4 at 2 p.m. If Sept 30 passes without a veto, only the AG can bring CIPA pen register claims over websites and apps from Jan 1. We talked regulatory risk vs class action exposure with Robert Freund on Deeper Diligence.

100

Daypart.ai @daypartai.bsky.social · 14/09/2026

Watch the full conversation, “AI Is Irrelevant. Whoever Has the Most Data Wins”: youtu.be/B7dtayeJQ-I?...

000

Daypart.ai @daypartai.bsky.social · 14/09/2026

Our latest podcast guest, Shoshana Wodinsky, walked us through one of the ugliest examples of what commercially available ad-tech data can do. A priest’s supposedly anonymous app data was used to figure out where he lived, which bars he visited and who he was, eventually costing him his job.

100

Daypart.ai @daypartai.bsky.social · 11/09/2026

Full ep: youtu.be/WyH88SVBp-U

000

Daypart.ai @daypartai.bsky.social · 11/09/2026

Steven Liss of OpenAds says getting an ad into an AI answer starts with getting the page retrieved, and that part is literally just SEO. Promptwatch has Reddit's ChatGPT citation share at 3.83% in late July and 0.52% by mid August.

100

Daypart.ai @daypartai.bsky.social · 10/09/2026

Location data keeps coming off military facilities after four DoD components turned off the MAID, and Wyden wants the IG to explain it. New episode with Shoshana Wodinsky on how it gets out. Same pipes your programmatic budget rides.

100

Daypart.ai @daypartai.bsky.social · 09/09/2026

Link to the latest Deeper Diligence episode here: youtu.be/NirjAy-0U5g?...

000

Daypart.ai @daypartai.bsky.social · 09/09/2026

Two payment processors in six days. Nuvei on September 4 for $4.85 million. Humboldt yesterday for $12 million plus a permanent ban on high risk merchant categories. We asked Robert Freund whether the enforcement picture looks the way most people assume it does.

100

Daypart.ai @daypartai.bsky.social · 08/09/2026

A Pennsylvania bill would make advertisers disclose when AI generated or substantially modified an ad. It cleared the House 124-78 and is with a Senate committee that returns September 28. We asked @RobertFreundLaw what that changes. New Deeper Diligence is up on our YouTube.

100

Daypart.ai @daypartai.bsky.social · 04/09/2026

Full newsletter: newsletter.daypart.ai/p/internatio...

000

Daypart.ai @daypartai.bsky.social · 04/09/2026

Before the next media plan goes out, check who can actually sue, whether you get a chance to fix it and if the vendor still has your back. In Colorado, that last part can disappear if the violation is yours.

100

Daypart.ai @daypartai.bsky.social · 04/09/2026

The EU runs two tracks at once. Regulators can fine under the AI Act, while consumers can bring claims under the revised Product Liability Directive.

110

Daypart.ai @daypartai.bsky.social · 04/09/2026

BIPA dates to 2008, back when the use case was fingerprint scanners, and it is still doing more damage in advertising than anything with AI in the title. Tennessee’s ELVIS Act works through a publicity law from 1984.

100

Daypart.ai @daypartai.bsky.social · 04/09/2026

A private right of action changes who has an incentive to care. An attorney general has a docket and a cure window. A plaintiff’s lawyer has a potential class and a fee award. Illinois sets damages at $1,000 for a negligent violation and $5,000 for a reckless one, per person.

100

Daypart.ai @daypartai.bsky.social · 04/09/2026

Texas and Colorado leave enforcement to one state office. Illinois and Tennessee let the person claiming harm file themselves.

100

Daypart.ai @daypartai.bsky.social · 04/09/2026

Last week we looked at what happens when AI creative invents a claim, borrows a face or shows up in a contract nobody updated. Part two follows up with one question: who gets to sue you?

100

Daypart.ai @daypartai.bsky.social · 04/09/2026

Full episode here: youtu.be/NirjAy-0U5g?...

000

Daypart.ai @daypartai.bsky.social · 04/09/2026

Sony Music sued Kroger last month over at least 392 alleged unlicensed uses of its recordings in social ads, brand accounts and paid influencers both. We asked @RobertFreundLaw for the worst case scenario.

100

Daypart.ai @daypartai.bsky.social · 04/09/2026

Adweek wrote this week about the ad industry running into the less glamorous side of AI. Imperfect performance, limited controls, no answer on accountability. On the latest Deeper Diligence we're joined by @RobertFreundLaw on where the exposure lands.

000

Daypart.ai @daypartai.bsky.social · 03/09/2026

Full episode here: youtu.be/NirjAy-0U5g?...

000

Daypart.ai @daypartai.bsky.social · 03/09/2026

On Monday Amazon got sued by the FTC with 22 states over it's ad auction pricing. Amazon answered back with performance data. On the latest episode of Deeper Diligence we are joined by @RobertFreundLaw who discusses what the exposure could look like.

100

Daypart.ai @daypartai.bsky.social · 01/09/2026

Full ep: youtu.be/WyH88SVBp-U

000

Daypart.ai @daypartai.bsky.social · 01/09/2026

Steven Liss of OpenAds calls web 1.0 advertising sort of a dumpster fire, constant banners and pop-up ads. In 2026 ai agents come with an inherent ad blocking mechanism, since they only retrieve what they think answers the query.

100

Daypart.ai @daypartai.bsky.social · 31/08/2026

Steven Liss thinks the LLM companies have an incentive to protect users here, down to prompting agents not to retrieve unlabeled sponsored information. He also thinks someone in ad tech goes after prompt injection poisoning anyway.

100

Daypart.ai @daypartai.bsky.social · 28/08/2026

Full ep: youtu.be/WyH88SVBp-U

000

Daypart.ai @daypartai.bsky.social · 28/08/2026

OpenAI sells ChatGPT ads on impressions or clicks. Steven Liss told us advertisers would rather price it on cost per action, since "advertiser's favorite currency is currency." AdExchanger this morning on how ChatGPT justifies its CPMs.

100

Daypart.ai @daypartai.bsky.social · 28/08/2026

Yesterday’s newsletter has the full version, from the contracts behind AI-generated posts to the reputational mess once the comments decide something is slop. daypart.beehiiv.com/p/generative...

000

Daypart.ai @daypartai.bsky.social · 28/08/2026

The source card includes the FTC’s joint AI statement and Endorsement Guides, New York’s synthetic-performer law, the U.S. Copyright Office report, the Pangram/YouGov survey and a Holland & Knight analysis.

100

Daypart.ai @daypartai.bsky.social · 28/08/2026

Generative AI adds new checks to even a strict agency review process. A nonexistent product feature or an AI-generated person can look plausible at first glance, so both need to be caught before the post receives final approval.

100

Daypart.ai @daypartai.bsky.social · 28/08/2026

YouGov surveyed 2,557 U.S. adults for Pangram. Among respondents who consume online content, 67% said they had seen AI content they believed was misleading or incorrect. They may have misidentified it, and by then a company account may already be engaging with consumers in the comments section.

100

Daypart.ai @daypartai.bsky.social · 28/08/2026

A sponsored post may need an ad disclosure. In New York, knowingly using an AI generated character in advertising can trigger another disclosure. Copyright is another question. Protection depends on what a human actually contributed, case by case.

100

Daypart.ai @daypartai.bsky.social · 28/08/2026

The tool warns its output may be wrong, the agency contract promises original and cleared work (procurement approved something in between), and the company’s name stays on the post, which is roughly when responsibility starts getting passed around.

100

Daypart.ai @daypartai.bsky.social · 28/08/2026

The FTC said there is “no AI exemption to the laws on the books.” If an AI-generated post invents a certification or depicts a feature the product does not have, the company still published the claim.

100

Daypart.ai @daypartai.bsky.social · 28/08/2026

Our latest newsletter is about the liability attached to AI-generated posts.

100