Couples are using AI to fight — and win — arguments: 'ChatGPT'
ChatGPT can be used for virtually anything, and in the world of romance, people have used the artificial intelligence tool to plan their weddings, write their wedding vows, score matches on Tinder, and find online companionship in the form of AI girlfriends — even if they’re happily married.
Using ChatGPT for Arguments
Now couples are going further than using AI to nourish their relationships — they’re using ChatGPT to help them win arguments in a fight. One person took to Reddit to share that their girlfriend uses the platform “every time we have a disagreement.”
“AITAH [Am I The A–hole] for saying she needs to stop?” the person asked in the AITAH subreddit.
The 25-year-old man explained that he and his 28-year-old girlfriend have had a couple of big arguments as well as a few smaller disagreements in their eight months of dating. And every time they have a disagreement, the girlfriend “will go away and discuss the argument” with ChatGPT — even sometimes while still in the same room.
He continued to say that when she does this, “she’ll then come back with a well-constructed argument breaking down everything I said or did during our argument.”
“I’ve explained to her that I don’t like her doing so as it can feel like I’m being ambushed with thoughts and opinions from a robot. It’s nearly impossible for a human being to remember every small detail and break it down bit by bit but AI has no issue doing so,” the user wrote.
Concerns and Discussions
He said that whenever he expresses his feelings about the situation, he’s told by his girlfriend that “ChatGPT says you’re insecure” or “ChatGPT says you don’t have the emotional bandwidth to understand what I’m saying.”
“My big issue is it’s her formulating the prompts so if she explains that I’m in the wrong, it’s going to agree without me having a chance to explain things,” he wrote.
Many people in the comments agreed with the user, noting that ChatGPT is “biased to the user’s input.”
“It’s literally programmed to tell you exactly what you want to hear. Discuss her actions with ChatGPT from your perspective and it’ll do the exact same thing to her,” the commenter said. “Show her how it’s biased and only serves as an artificial form of self-validation.”
Someone else added, “I noticed that…it is programmed to reinforce your position. It’s machine learning to an absurd degree, but still machine learning. It asks people to rate the responses. She thinks it’s impartial because it’s a robot, but it’s a robot programmed to tell people what they want to hear.”
One user even put the man’s situation back into ChatGPT to ask if they were indeed the a–hole in the situation, and ChatGPT itself said: “While AI can be helpful for many things, it shouldn’t replace genuine, human-to-human conversations that are nuanced, emotional, and require empathy…While AI can provide thoughtful input, it’s not a substitute for emotional intelligence and understanding the complexity of relationships.”
The AI tool also noted that, “As you mentioned, the way she frames her prompts affects the advice or feedback she gets. If she primarily explains the situation in a way that favors her side, the response will likely mirror that. This makes it a one-sided tool rather than a fair mediator.”
Others joked that the man should tell his girlfriend that ChatGPT told him he should break up with her.
“Respond with ChatGPT until she gets the point,” someone quipped.
“Tell her you consulted with ChatGPT and it told you to break up with her,” another said.
“NTA. Tell her to put this prompt in: ‘How do I break up with my girlfriend?'” a user joked.
Recent Developments in AI
This week in AI, synthetic data rose to prominence. OpenAI last Thursday introduced Canvas, a new way to interact with ChatGPT, its AI-powered chatbot platform. Users can generate text or code in Canvas, then, if necessary, highlight sections to edit using ChatGPT.
A new feature called Canvas lets ChatGPT users focus on specific lines for collaboration with the AI interface. OpenAI and Microsoft are funding projects to bring more AI tools into the newsroom.
Anthropic's latest development gives its Claude AI tool the ability to control a PC, reportedly just like a person would. One of the only U.S. government offices dedicated to assessing AI safety is in danger of being dismantled if Congress doesn't choose to authorize it. The U.S. AI Safety Institute (AISI), a federal government body that studies risks in AI systems, was created in November 2023 as a part of President Joe Biden's AI Executive Order. The AISI operates within NIST, an agency of the Commerce Department that develops guidance for the deployment of various categories of technologies.
The open-source AI image generator Stable Diffusion has been updated to version 3.5. It addresses previous issues, improves performance with prompts, and offers greater skin tone and feature diversity.
What do the actor Kevin Bacon, the novelist Kazuo Ishiguro, the musician Robert Smith, and the journalist and historian Sidney Blumenthal have in common? “The unlicensed use of creative works for training generative AI is a major, unjust threat to the livelihoods of the people behind those works and must not be permitted,” signatories warn.
Cash App is a peer-to-peer payment service that also offers business transactions and banking services. Here's what to know about how it works and whether it's safe to use.
Startups that build AI medical assistants and medical scribes, which save doctors time that they would otherwise spend taking notes and populating medical records, have been booming over the past year. On Tuesday, video-conferencing company Zoom said it is partnering with Suki, one such AI medical scribe provider, to offer doctors on its platform an AI scribe that can take notes about their consultations with patients. Zoom is used for about 36% of all telehealth visits in the U.S., making it the most popular video-conferencing platform, according to industry intelligence group Definitive Healthcare.
Google Messages adds new safety features, including enhanced scam detection. There’s also a forthcoming tool to disable texts from unknown international senders.
