ChatGPT doesn't require an account – but there's a catch | Tech Crunch

Spread the love

OpenAI is making its flagship conversational AI available to everyone, even people who don't bother creating an account. But it's not the same experience – and all your chats go into their training data unless you opt out.

Starting today in select markets and gradually expanding to other parts of the world, visiting will no longer ask you to log in — although you can if you want to. Instead, you'll be dropped into a conversation with ChatGPT, which uses the same model as logged-in users.

You can chat to your heart's content, but keep in mind that you won't get the same set of features that people with accounts have. You may not save or share chats, use custom suggestions, or use other features normally associated with a persistent account.

One suspects that you still have the option to opt out of your chats being used for training (which undermines the whole reason the company is doing it in the first place). Click the little question mark in the lower right, then click “Settings” and disable the feature there. OpenAI has this useful gif:

Image Credits: OpenAI

More importantly, this extra-free version of ChatGPT has “slightly more restrictive content policies.” What does that mean? I asked and got a verbose yet largely nonsensical answer from a representative:

The signed-out experience benefits from existing security mitigations already built into the model, such as refusing to generate harmful content. In addition to these existing mitigations, we are also implementing additional safeguards specifically designed to address other types of content that are inappropriate for the signed-out experience.

Informed by our understanding of the GPT-3.5 capabilities and risk assessments we've completed, we've considered potential ways the Logged Out Service could be used in inappropriate ways.

So… really, no clue what these more restrictive policies are. No doubt we'll soon find out as an avalanche of randos descend on the site to kick the tires on this new offering. “We recognize that additional iteration is necessary and welcome feedback,” the spokesperson said. And they have it – in abundance!

I also asked if, by that time, they had any plans for how to handle what would almost certainly be attempts to abuse and weaponize the model on an unprecedented scale. Just think: a platform where a billionaire can lose money. After all, inference is still expensive, and even the refined, low-lift GPT-3.5 model takes up power and server space. People are going to hammer it for what it's worth.

They also have a non-verbal response to this threat:

We've also carefully considered how to detect and stop abuse of the Signed Out experience, and the teams responsible for detecting, preventing, and responding to abuse have been involved in the design and implementation of this experience and will continue to inform its design. Moving forward.

Note the absence of anything resembling specific information. People have little idea that they are going to inflict this on anyone else and need to be reactive rather than proactive.

It's not clear which regions or groups will get access to the ultra-free ChatGPT first, but it starts today, so check back regularly to see if you're among the lucky ones.

Source link

Leave a Comment