ChatGPT is ‘getting more seductive’ as free upgrade adds voice compared to ‘flirty virtual assistant’ in the movie Her
ARTIFICIAL Intelligence (AI) company OpenAI has added a new flirty feature to its ChatGPT app.
ChatGPT is adding a voice feature that not only allows the user to speak directly to the AI but also allows the AI bot to talk back.
The AI bot chatting program launched on app stores in November of 2022, and was met with instant success.
What is alluring about the app is that people can ask it questions and gain a response that is in the form of typical human communication.
This differs from a Google search where users have to sort through information themselves, as they can customize response settings.
“These models were trained on vast amounts of data from the internet written by humans, including conversations, so the responses it provides may sound human-like,” Open AI explained.
Because the AI learns based on interactions with other users in addition to data fed to it by the company and open-access information on the internet, its answers are not always accurate.
“ChatGPT is not connected to the internet, and it can occasionally produce incorrect answers. It has limited knowledge of world and events after 2021 and may also occasionally produce harmful instructions or biased content,” warned the app’s parent company.
For instance, when asked by the U.S. Sun about its new speaking feature, the chat responded that no such feature existed.
“I apologize for any confusion, but I don’t have access to real-time information or updates beyond my last training data in September 2021. As of my last update, I do not have the capability to generate speech or engage in voice communication,” responded the AI bot.
As the app has experienced continued success, the company has been working on updates, now launching a feature just a few months shy of a one-year anniversary.
Similar to the AI assistant from the sci-fy movie Her, ChatCPT can now speak back to its user in a two-way conversation.
Conversations can range from instruction on how to cook to potentially flirty or seductive conversation, it all depends on what the user asks and feeds the bot.
Functionally, it first takes the user’s words and turns them into a transcript for the bot to read, then the bot’s response is “said” by a synthetically created voice.
This feature is only currently available for users with a $20-per-month subscription.
When WIRED tested a beta version of this new feature, they reported that there was some lag in the response.
This new feature still holds safeguards like the original version of the app by not answering dangerous requests such as 3D-printed gun parts or bomb building.
Although all versions of the app allow the user to bar ChatGPT from using their interactions as part of its learning, when this is turned off, WIRED reported that the new speaking feature no longer was functional.
According to the company, this is not because the user’s voice is used in training the bot.
“We currently don’t collect any voice data for training, and we are thinking about what we want to enable for users that do want to share their data,” explained Niko Felix, a spokesperson for OpenAI.
When it comes to the image feature of the app, users can opt in or out of their pictures being used for training.
“Users can opt out of having their image data used for training. Once opted-out, new conversations will not be used to train our models,” said Felix.
New users have mixed reactions to the new conversation feature.
“It’s not someone and you’re not even talking with something. You’re talking at a machine which is generating sentences that you imbue with meaning,” wrote one X user in response to the idea of talking to Chat GPT therapeutically.
“A conversation requires thinking things on both sides, but what chatgpt does isn’t thinking,” they finished.
“We are all in a simulation,” posted another user.
“That’s an exciting development!” expressed a different post.