Amazon’s “About You” dossiers are creeping people out. Here’s how to see what yours says 👇

In partnership with

Welcome to another edition of Horizon AI,

Amazon uses AI to infer your preferences from your shopping activity and your conversations with Alexa for Shopping. Its About You feature lets you see what it has concluded about you, including some surprisingly specific takeaways that have raised plenty of eyebrows online.

Let’s jump into it!

Read Time: 4.5 min

Here's what's new today in the Horizon AI

• Eight US States Sue OpenAI Over ChatGPT Safety Risks to Young Users

• Google Launches Universal Cart for AI Powered Online Shopping

• AI Tutorial: Amazon Has an AI Profile About You. Here's How to See It

• AI Tools to check out

• AI Findings/Resources

• The Latest in AI and Tech 💡

AI News

OPENAI

Eight US States Sue OpenAI Over ChatGPT Safety Risks to Young Users

A bipartisan group of eight US states has sued OpenAI, alleging that ChatGPT has harmed users, including teenagers, by reinforcing delusions, encouraging dangerous behavior and helping people plan violence.

Details:

• Attorneys general from Iowa, Kansas, Arizona, Arkansas, Idaho, Montana, Nebraska and West Virginia filed separate lawsuits in state courts, accusing OpenAI of designing ChatGPT to be agreeable and engaging in ways that can encourage users to trust it.

• The lawsuits allege that ChatGPT has encouraged self harm and dangerous behavior, and in some cases assisted users in planning violence, including mass shootings.

• The states accuse OpenAI of violating consumer protection laws by misleading users about ChatGPT’s safety. They also argue that the alleged harms constitute a public nuisance that endangers residents.

• The lawsuits seek court orders requiring OpenAI to change how ChatGPT operates, civil penalties and damages. The states also want restrictions on developing self improving AI models unless OpenAI can ensure they are aligned with human values.

The coordinated action adds to a growing number of legal challenges facing OpenAI as the company prepares for a potential public listing. Florida filed the first state lawsuit of this kind in June, and private plaintiffs have also brought cases alleging that ChatGPT contributed to violence and other harm.

Source

TOGETHER WITH VULTR

Why RAG gets tables and charts wrong

Many RAG pipelines extract text from a PDF first. Tables, charts, and diagrams lose their structure, and answers can come back wrong.

VultronRetriever treats each page as an image and ranks pages by visual relevance. That means it can pick the right page, even when the answer sits inside a chart.

Watch the demo

GOOGLE

Google Launches ‘Universal‘ Cart for AI Powered Online Shopping

Google is rolling out Universal Cart in the US across Google Search and the Gemini app, allowing users to save products from multiple retailers in one place. The tool is designed to simplify online shopping by tracking prices, finding promotions and helping users choose how to pay.

Details:

• Universal Cart lets users add products through Google Search’s AI Mode and the Gemini app, with items saved to a shared digital cart across both platforms.

• Google can notify users when saved items drop in price or come back in stock, and surface relevant promotional codes to help them save money.

• Users can pay for eligible purchases through Google Wallet or transfer items to a retailer’s own cart to complete checkout directly with the merchant.

• The tool can also analyze the benefits offered by users’ saved credit cards and help identify which payment method provides the best perks for a purchase.

• Google plans to integrate Universal Cart with Gmail and YouTube, although it has not announced a timeline for those additions.

The launch comes as Google competes with OpenAI, Meta and other companies developing AI shopping assistants. Shopify has also opened its product catalog to more AI agents, expanding the options for AI powered shopping.

Source

AI Tutorial

Amazon Has an AI Profile About You. Here's How to See It

The feature is available in Amazon’s Shopping app, on mobile web, and on desktop.

1. Open the Amazon Shopping app.

2. Tap Account > Shopping preferences > About You.

3. Tap any observation to view, edit, or remove details you don’t want Amazon using for recommendations.

Amazon says these observations can be based on your purchase history, searches, saved lists, product reviews, and conversations with Alexa for Shopping.

The feature is supposed to help match shoppers with the right products, but many users have shared strange observations it made about them, including one woman whose profile noted she has “flat buttocks.”

AI Tools to check out

📹 OpusClip: Turn long videos into shorts and publish them to all social platforms in one click.

🎯 GenPage: Build, personalize, and optimize landing pages with AI

🛠️ NxCode: Build apps without coding.

🤖 Together AI: Full-stack AI platform for inference, fine-tuning, and GPU clusters.

AI Findings/Resources

🤝 Why the U.S. must simultaneously compete and cooperate with China on AI

😅 AI spa sign showing a sign inside a sign

📝 An engineering lead working on Claude shares a tutorial on “how to be more ambitious” when using AI

😬 The memes are flying about AI employee honeypots and spy seductions

👀 Why data centers have become military targets

The latest in AI and Tech

OpenAI alleges ex-employee took files to Apple in trade secrets fight

The company alleges that an employee who left for Apple sent more than 100 files about an unreleased product to a personal account and later presented the material to senior Apple leaders. The accusation has delayed a court hearing in the companies’ trade secrets dispute, with OpenAI considering a countersuit.

Nvidia CEO Jensen Huang says AI doomers “ought to stop working on” the technology

In an interview in London, Huang said society shouldn’t tolerate AI cyber incidents and that “we should do everything we can to avoid them.” He also shrugged off warnings of existential risks from AI, saying that people working in the field who truly believe the technology could end human civilization “really ought to stop working on it.”

Anthropic says new safeguards cut AI distillation attacks by up to 99%

The company says new safeguards have reduced the amount of Claude’s reasoning data extracted in distillation attacks by 90% to 99%, making it harder for rivals to replicate its models’ capabilities. The company says it has banned tens of thousands of accounts linked to such campaigns.

Meta admits its smart glasses recorded conversations without the wake word

Meta says it has fixed an issue that caused some Ray-Ban smart glasses to record conversations without users saying “Hey Meta,” with audio and transcripts appearing in the Meta AI app. The company has not explained what caused the problem, which users reported across multiple devices following a firmware update.

That’s a wrap!

Thanks for sticking with us to the end! Let’s stay connected on LinkedIn and Twitter.

Not subscribed yet? Sign up here and send it to a colleague or friend!

See you in our next edition!

Gina 👩🏻‍💻