Starbucks Plans 2027 Drive-Thru AI as GM Pays $220M Over Car Data

Starbucks plans to test artificial-intelligence systems at its drive-thrus starting in 2027 that recognize arriving customers, display personalized digital menus, and assist baristas with order entry, according to The Verge, even as General Motors agreed to pay $220 million to settle a 36-state lawsuit over nonconsensual connected-car telemetry sales (KCBS Radio), online fashion retailer Revolve disclosed an automated payment-card breach (The Record), and network research firm Graphika detailed China-linked Spamouflage accounts posing as U.S. voters on X (KCBS Radio).

Starbucks Plans 2027 Drive-Thru AI and App Search Upgrades

During Thursday's earnings call, Starbucks CEO Brian Niccol said the coffee chain is exploring drive-thru systems that recognize participating customers upon arrival, present personalized digital menu boards, and streamline payment, alongside background AI models launching in 2027 to help baristas input orders more accurately and efficiently.

Starbucks also plans to add AI-powered search, discovery, and conversational or agentic ordering features to the Starbucks mobile app, expanding on its existing integration allowing customers to place orders through OpenAI's ChatGPT and its employee-facing "Green Dot Assist" assistant for recipe and service-standard queries. The drive-thru initiative parallels Taco Bell's car-by-car digital menu board experiments and McDonald's pilot of its "ArchIQ" chatbot for repeat drive-thru customers.

General Motors Pays $220 Million to Settle 36-State OnStar Telemetry Lawsuit

General Motors will pay $220 million to settle allegations brought by a coalition of 36 state attorneys general — led by Oregon Attorney General Dan Rayfield and the Washington Attorney General — that the automaker collected and sold drivers' geolocation and driving-behavior telemetry to data brokers and insurance companies without informed consent.

According to Rayfield's office, GM enrolled millions of buyers in its Smart Driver program between 2016 and 2024 during dealership OnStar sign-ups, where sales staff earned enrollment bonuses and dealerships were ranked by enrollment volume. Investigators found that dealership employees frequently accepted terms on customers' behalf without permission, and that more than 370,000 drivers were erroneously told in the OnStar app that they were not enrolled in Smart Driver even as their trip times, hard braking, acceleration, and speed data were sold to brokers for insurance rate calculations (at a time when LendingTree pegs average 2026 U.S. auto insurance at $177 per month or $2,124 per year for full coverage, and $68 per month or $816 per year for minimum coverage). Following an earlier Federal Trade Commission (FTC) consent order barring GM and OnStar from sharing telemetry with consumer reporting agencies, the 36-state settlement requires GM to obtain affirmative, express consent before collecting or sharing driving or location data, limit data collection, and implement dealership compliance oversight and external privacy audits.

Revolve SEC 8-K Payment Card Breach and Graphika's Spamouflage Report

In cybersecurity and online influence disclosures released Friday:

• Revolve Payment Card Breach: In a Thursday Form 8-K filing with the Securities and Exchange Commission (SEC), online fashion retailer Revolve — founded in 2023, carrying more than 1,600 brands, and reporting $347.4 million in net sales in its most recent fiscal quarter — warned that attackers used "automated means" to access saved customer names, contact details, payment card numbers, and expiration dates on its e-commerce platform. Revolve notified federal law enforcement, retained an external cybersecurity firm, and reported to state regulators on Friday that 127,200 residents in Texas and 1,328 residents in Vermont had data stolen, alongside breach filings in Massachusetts.
• Graphika Flags Spamouflage Accounts on X: New York-based network analysis firm Graphika reported Friday that accounts linked to China's "Spamouflage" operation are posing as U.S. voters on X ahead of Tuesday's midterm elections to question election integrity and U.S. democratic legitimacy, including 4 accounts amplifying political personas and another account with more than 200,000 followers repurposed from a Philippine culture blog that posted attacks on Democratic Party leadership during the October 2025 government shutdown. Darren Linvill, co-director of Clemson University's Watt Family Innovation Center Media Forensics Hub, noted that while Spamouflage accounts are ineffective at mimicking American voters, they frequently target and harass members of the Chinese diaspora community.