AI News Brief for Thursday, May 14, 2026

  • When AI Harms You, Who's Responsible?
    States are moving to close a legal gap that leaves consumers with no clear path to recourse when an AI system causes real harm.

  • Ten Text Scams That Are Getting Harder to Spot.
    AI is giving scammers a significant upgrade. Here's what to watch for in your inbox and how to protect yourself.

When AI Harms You, Who's Responsible?

Policy & Regulation
Source: NPR
https://www.npr.org/2026/05/11/nx-s1-5798754/several-states-considering-ban-on-legal-personhood-for-ai

In a nutshell

Several U.S. states are drafting legislation to explicitly bar AI from being granted legal personhood, preempting potential arguments that AI systems should bear legal responsibility for harmful actions. The debate centers on accountability gaps when AI causes harm.

Why it's important

These bills aim to keep accountability squarely with human actors and companies, which is critical for consumers who want clear recourse when AI errors affect medical decisions, financial advice, or fraud protection.

Key takeaways

  • Several U.S. states, including Ohio, are drafting bills to ban any form of legal personhood for AI, ensuring that a human or company always stands legally responsible when an AI causes harm.

  • Under current law, when an AI makes a harmful decision, only the company that built it can be sued. As AI takes on more independent actions, legal experts say that framework has a serious gap.

  • A Florida mass shooting case put the issue in focus. Prosecutors noted that "if it was a person on the other end of that screen, we would be charging them with murder," but existing law offers no path to hold an AI accountable.

  • Legal experts describe a genuine "hole in the law" as AI increasingly acts as an independent agent in areas like financial management, healthcare planning, and legal paperwork.

  • Unlike traditional software, AI systems are "trained" rather than "programmed," meaning their behavior cannot be fully predicted. As one researcher put it: "AI is basically like [a trained dog]. It can choose to do something else, and sometimes it will."

  • State-level bills banning AI legal personhood are designed as a consumer protection measure, ensuring you always have a human or company to hold accountable if an AI tool causes real harm.

Your thoughts?

If an AI tool you relied on made a financial or medical decision that caused you harm, who do you think should be responsible for making it right?

Ten Text Scams That Are Getting Harder to Spot

Privacy & Data Security
Source: Saving Advice
https://www.savingadvice.com/articles/2026/05/07/10734312_10-toll-text-and-amazon-scams-exploding-in-2026.html

In a nutshell

AI is making text-based scams significantly harder to detect, with criminals using AI tools to generate professional-sounding, personalized messages mimicking toll authorities and delivery services. Older Americans report receiving dozens of these scam texts weekly, and AI allows scammers to scale attacks and tailor messages using data from prior breaches.

Why it's important

Adults 50+ are disproportionately targeted by AI-enhanced text scams; the article identifies the most common patterns and provides practical guidance on what to watch for.

Key takeaways

  • Fake toll payment texts are one of the fastest-growing scam types, claiming you owe small amounts like $3.95 to create urgency. Consumer Reports says virtually every toll-payment text demanding immediate payment is fraudulent.

  • Amazon-related scams now include realistic order numbers, cloned logos, and fake recall notices that are nearly indistinguishable from genuine customer service alerts.

  • AI tools allow scammers to write grammatically perfect, personalized messages using data from past breaches, including real names and partial account numbers, making these texts far harder to identify as fraud than they were even a year ago.

  • The small payment request is rarely the real goal. Scammers use it to capture credit card numbers, passwords, and identity details for larger fraud schemes later.

  • Adults 50+ are targeted aggressively because scammers know many retirees rely on text alerts for deliveries, medications, and account notifications, and are more likely to act quickly when a message feels urgent. Some are receiving dozens of these texts every week.

  • The safest habit is simple: never click links in unexpected texts, even when the message looks official. Forward scam texts to 7726 (SPAM) to help carriers block them, and report to ReportFraud.ftc.gov.

Your thoughts?

When you get an unexpected text about a package, payment, or account problem, what's your process for deciding whether it's real before you do anything?

This brief was prepared by Martin Danner (a human) using Claude Cowork as a research assistant.