Tesla Under Investigation, Anthropic Under Fire

AI hits the streets (and the courts) with major consequences.

In partnership with

Wednesday Deep Dive

(Reading Time: 4 minutes)

Find out why 1M+ professionals read Superhuman AI daily.

In 2 years you will be working for AI

Or an AI will be working for you

Here's how you can future-proof yourself:

  1. Join the Superhuman AI newsletter – read by 1M+ people at top companies

  2. Master AI tools, tutorials, and news in just 3 minutes a day

  3. Become 10X more productive using AI

Join 1,000,000+ pros at companies like Google, Meta, and Amazon that are using AI to get ahead.

The Wednesday Deep Dive takes a detailed look at what's new in AI. Each week, we share in-depth insights on new tools, proven prompts, and significant developments - helping tech professionals work smarter and stay ahead.

This week’s Deep Dive tracks two stories at the edge of AI’s regulatory frontlines: from courtroom copyright fights to AVs drifting through the wrong lane in Texas.

🔍 Anthropic wins a landmark fair use ruling (but still faces a pirate book trial)
🚗 Tesla’s robotaxis go live, and immediately attract federal scrutiny

Let's dive in.

🌐 AI News

🧠 Anthropic scores a big win on book training, but it's not out of the woods yet

A federal judge just handed Anthropic a major legal victory that could ripple through the AI industry: training AI on legally purchased physical books is considered fair use under U.S. copyright law.

The case, brought by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, accused Anthropic of training Claude models on copyrighted books without permission. But Judge William Alsup ruled that when Anthropic bought printed books, scanned them, and used the resulting text to train its models, that act met the criteria for fair use.

“Authors’ complaint is no different than it would be if they complained that training schoolchildren to write well would result in an explosion of competing works,” Alsup wrote.

🧾 What the ruling does say:

  • Purchasing physical books and scanning them is a lawful practice.

  • Using those scans to train a large language model is sufficiently transformative to qualify as fair use.

  • The Copyright Act is designed to advance original work, not shield authors from competition.

📕 What the ruling doesn’t say:

  • Pirated content remains in limbo; a separate trial will decide that question.

  • The decision is silent on whether AI-generated outputs infringe copyright, a key issue in other cases.

  • It also sets no precedent for training on ebooks or licensed text scraped from the web.

Judge Alsup made it clear that a separate trial will proceed for Anthropic’s alleged use of pirated books that were downloaded from illegal sources.

🤔 Why It Matters:

This is the first ruling to declare that training on scanned books you legally own is fair use, which could shield AI developers who follow similar protocols. But the win is narrow, and the ruling makes a sharp distinction between lawful training sets and digital piracy.

Expect more cases to split this legal hair.

🌐 AI News

🚗 Tesla robotaxis hit the streets, and regulators hit the brakes

On Sunday, Tesla officially launched a limited test of its long-promised robotaxi service in Austin, Texas, marking the company’s first real-world deployment of fully driverless cars. But by Monday, videos of troubling behavior (including driving in the wrong lane and speeding past another autonomous vehicle) had already triggered a response from federal safety regulators.

The National Highway Traffic Safety Administration (NHTSA) confirmed it is reviewing the videos and has contacted Tesla for more information. In a statement, the agency said it was “aware of the referenced incidents and is in contact with the manufacturer to gather additional information.”

Tesla responded that its safety documentation is “confidential business information” and has asked that responses to the agency’s questions remain private.

🚕 What Tesla Deployed:

  • Around a dozen driverless vehicles are now operating in Austin as part of a paid, limited-use robotaxi test.

  • Users must be 18 or older to ride.

  • Vehicles avoid bad weather conditions and difficult intersections, raising questions about the reliability of routes.

  • The service runs on Tesla’s Full Self-Driving (FSD) software stack, despite that system still being in beta and marketed with disclaimers.

This is a pivotal moment for Tesla, whose robotaxi promises date back to 2019, when Elon Musk claimed the company would have “a million robotaxis on the road” by 2020. That vision is now partially realized, but under tight constraints and increasing scrutiny.

📉 NHTSA's Broader Investigation:

Tesla’s AV tech has been under formal federal investigation since October 2023, following a string of collisions involving vehicles using FSD in low-visibility or nighttime conditions. The current probe covers an estimated 2.4 million Tesla vehicles and focuses on whether the system meets basic road safety expectations.

In one of the most serious incidents, a 2023 crash involving an FSD-equipped Tesla resulted in a fatality, prompting lawmakers and researchers to demand stronger oversight.

Importantly, NHTSA clarified that it does not “pre-approve” new driving technologies. Automakers certify compliance themselves, but the agency can still investigate incidents and demand recalls if safety defects are found.

🛣️ The Bigger AV Picture:

Tesla isn’t alone in testing AV fleets. Companies like Waymo, Cruise, and Zoox have been running pilot programs for years. But Tesla’s approach is unique:

  • It skips LIDAR and external sensors in favor of a vision-only system.

  • Its software is tightly integrated with the company’s consumer vehicles, not bespoke robotaxis.

  • It launched without third-party safety validation or external supervision.

This approach has drawn both praise for its ambition and criticism for its risk.

Unlike competitors who require remote monitoring and conservative routing, Tesla’s AVs rely entirely on internal AI decision-making. That’s bold, but if early signs of lane drifting and erratic behavior continue, it could backfire.

🚦Why It Matters:

Tesla’s rollout is a milestone and a stress test for U.S. safety frameworks. If NHTSA concludes that Tesla’s robotaxis present meaningful risks to pedestrians, passengers, or other drivers, the agency could force changes or even halt the pilot.

This moment also tests the credibility of Tesla’s core AV thesis: that its cars can safely operate without expensive sensors, human backup, or external verification.

So far, the real-world footage has raised more questions than answers. Now the regulators are watching.

Want to Unsubscribe from future emails? Just hit “unsubscribe” below.