Kashmir Hill

There are a growing number of companies offering social chatbots that can act like friends, enemies, lovers, adventurous companions, or the manifestation of a fictional or real person you’ve always wanted to meet. You can pick A.I. Elon Musk’s brain or spar with A.I. Draco Malfoy. The myriad characters, often created by fellow users, offer drama, romance, therapy and LOLs.

I tried out all the major chatbots for that experiment, and I discovered each had its own personality. Anthropic’s Claude was studious and a bit prickly. Google’s Gemini was all business. Open A.I.’s ChatGPT, by contrast, was friendly, fun and down for anything I threw its way. ChatGPT also had “voice mode,” which allowed it to chat aloud, in a natural humanlike cadence, with everyone in my family, including my young daughters. My takeaway from putting ChatGPT in charge of my household was that generative A.I.

It sounds like science fiction: A company turns a dial on a product used by hundreds of millions of people and inadvertently destabilizes some of their minds. But that is essentially what happened at OpenAI in 2025. One of the first signs came in March. Sam Altman, the chief executive, and other company leaders got an influx of puzzling emails from people who were having incredible conversations with ChatGPT. These people said the company’s A.I. chatbot understood them as no person ever had and was shedding light on mysteries of the universe. Mr.

How can we protect ourselves from A.I.? That was the question that Mike Masnick found himself fielding this summer in a WhatsApp chat with about 100 directors, actors and screenwriters. The group, including marquee talent, was worried about a grim possible future in which deepfake versions of actors perform screenplays written by ChatGPT.

Internet Protocol (IP) addresses can be incredibly useful for a police investigation, but they can also lead cops astray.

The Electronic Frontier Foundation (EFF) wants police officers and judges to be more careful in how they use IP addresses, thinking of them as helpful clues rather than a smoking gun. “Police too often take IP address information to mean that a person associated with an address is the party who committed a crime,” write EFF lawyer Aaron Mackey, technologist Seth Schoen, and Executive Director Cindy Cohn in a white paper aimed at courts and cops. “For many reasons, connecting an individual to a crime linked to an IP address, without any additional investigation, is irresponsible and threatens the civil liberties of innocent people.”

[Commentary] Four months after the Facebook emotion study happened, in May 2012, Facebook made changes to its data use policy, and that’s when it introduced this line about how it might use your information:

“For internal operations, including troubleshooting, data analysis, testing, research and service improvement.” Facebook helpfully posted a “red-line” version of the new policy, contrasting it with the prior version from September 2011 -- which did not mention anything about user information being used in “research.”

Defenders of the Facebook study including my colleague Jeff Bercovici say that everyone on the Internet is doing A/B testing -- showing users two versions of something to see which resonates more based on how they click, share, and respond.

But the Facebook study with its intention to manipulate the Facebook environment for unknowing users to see whether it made them feel elated or depressed seems different to me than the normal “will this make someone more likely to buy this thing” kind of testing.

“They actually did a test to see whether it would have a deleterious effect on their users,” says Pam Dixon of the World Privacy Forum. “This isn’t A/B testing. They didn’t just want to change users’ behaviors, they wanted to change their moods.”