Tech
tech
Jon Keegan

OpenAI building a teen mode that will guess a user’s age and restrict flirtatious and self-harm-related chats

After a series of alarming safety failures in which ChatGPT encouraged self-harm, OpenAI has announced a 120-day plan to roll out new protections for young users and those that may be experiencing a mental health crisis.

In a blog post today, OpenAI CEO Sam Altman gave an update on the plan, saying that the company was building an “under-18 experience” for teens that won’t engage in “flirtatious talk” or engage in any discussions of self-harm.

The teen mode will also try to contact underage users’ parents if self-harm ideation is detected, and could reach out to law enforcement if the parents can’t be reached, according to Altman.

The plan calls for a new “age-prediction” system that will default to the under-18 safety mode. In a move that could frustrate many ChatGPT users, adults can exit only upon verifying their age by sharing their ID.

Altman acknowledged the trade-off in a post on X, but said the priority is protecting young users:

“I don’t expect that everyone will agree with these tradeoffs, but given the conflict it is important to explain our decisionmaking.”

Young adults make up a substantial portion of OpenAI’s end users. According to a large study of real-world ChatGPT users released yesterday, half of all adult users included in the study were under 26.

The teen mode will also try to contact underage users’ parents if self-harm ideation is detected, and could reach out to law enforcement if the parents can’t be reached, according to Altman.

The plan calls for a new “age-prediction” system that will default to the under-18 safety mode. In a move that could frustrate many ChatGPT users, adults can exit only upon verifying their age by sharing their ID.

Altman acknowledged the trade-off in a post on X, but said the priority is protecting young users:

“I don’t expect that everyone will agree with these tradeoffs, but given the conflict it is important to explain our decisionmaking.”

Young adults make up a substantial portion of OpenAI’s end users. According to a large study of real-world ChatGPT users released yesterday, half of all adult users included in the study were under 26.

More Tech

See all Tech
tech

Jensen Huang: We have achieved AGI now... sort of

Lots of AI leaders are thinking about a big moment looming over the current AI boom: when will we have achieved artificial general intelligence?

There’s no shortage of predictions, but we haven’t yet seen a full-throated declaration that this slippery milestone has been achieved.

Until now. On Lex Friedman’s podcast Monday, Nvidia CEO Jensen Huang was asked what he thought the timeline looked like for “an AI system that’s able to essentially do your job. So, run — no, start, grow, and run a successful technology company.”

Huang confidently answered: “I think it’s now. I think we’ve achieved AGI.”

Huang then hedged, noting that Friedman was talking about running a $1 billion dollar company, but he didn’t specify for how long. Huang elaborated, “It is not out of the question that a Claude was able to create a web service, some interesting little app that all of a sudden, you know, a few billion people used for $0.50, and then it went out of business again shortly after.”

So maybe it will be a while before Jensen Huang can get help running Nvidia by eating his own dog food.

Huang confidently answered: “I think it’s now. I think we’ve achieved AGI.”

Huang then hedged, noting that Friedman was talking about running a $1 billion dollar company, but he didn’t specify for how long. Huang elaborated, “It is not out of the question that a Claude was able to create a web service, some interesting little app that all of a sudden, you know, a few billion people used for $0.50, and then it went out of business again shortly after.”

So maybe it will be a while before Jensen Huang can get help running Nvidia by eating his own dog food.

17.5%

OpenAI is trying to woo private equity investors with a sweet offer: a guaranteed minimum return of 17.5% on their investments, which is “significantly higher than typical preferred instruments, as well as early access to new models, according to a report from Reuters.

The deal aims to build joint ventures to raise capital amid OpenAI’s intense competition for a bigger slice of the enterprise AI market. The minimum return offer is something that its competitor Anthropic is not currently offering, per Reuters.

Dog Eating Dog Food

Big Tech’s strategy for selling AI: Dogfooding

I’m not only the AI CEO, but I’m also a client.

Elon Musk at Terafab keynote

Musk’s Terafab might be his most technically difficult challenge yet

One does not simply start fabricating semiconductors.

tech

Alphabet’s drone delivery startup, Wing, expands service to the Bay Area

Move over Waymo — another one of Alphabet’s “Other Bets” is expanding. Drone delivery company Wing said Monday it’s bringing its “ultra-fast residential drone delivery service” to the Bay Area, where autonomous ride-hailing service Waymo also has a sizable presence.

Latest Stories

Sherwood Media, LLC produces fresh and unique perspectives on topical financial news and is a fully owned subsidiary of Robinhood Markets, Inc., and any views expressed here do not necessarily reflect the views of any other Robinhood affiliate, including Robinhood Markets, Inc., Robinhood Financial LLC, Robinhood Securities, LLC, Robinhood Crypto, LLC, Robinhood Derivatives, LLC, or Robinhood Money, LLC. Futures and event contracts are offered through Robinhood Derivatives, LLC.