- A new chatbot called Pi, launched by Inflection AI, offers personalized advice and assistance.
- LinkedIn co-founder Reid Hoffman and DeepMind co-founder Mustafa Suleyman co-founded Inflection AI.
- Pi was designed to be user-friendly, but makes it clear to users that it cannot feel emotions.
There’s a new AI chatbot on the scene – and this one wants to get personal.
Pi, who means “personal intelligence”, was published tuesday by Inflection AI. The AI startup has three co-founders: Mustafa Suleyman, the co-founder of DeepMind; Karén Simonyan, the company’s chief scientist; and LinkedIn co-founder Reid Hoffman.
Pi is designed to be conversational and offer emotional support. In his opening dialogue, Pi tells users he’s meant to be “helpful, friendly, and fun” and invites users to ask him for advice or answers, or talk about anything that’s on their mind.
Suleiman told Bloomberg Pi is meant to be a “neutral listener” who asks follow-up questions and offers advice.
When I told Pi “life has been overwhelming”, the chatbot probed me to share more.
I told Pi I felt overwhelmed and the chatbot responded with several questions.
Pi/AI Inflection
“A lot of people feel like they just want to be heard and they just want a tool that reflects what they said to demonstrate that they were actually heard,” Suleyman told Bloomberg.
Like OpenAI’s ChatGPT and Google’s Bard chatbots, Pi is based on LLMs, or large language models. This means that bots are trained to generate text responses by analyzing large amounts of content available online.
In another conversation, I told Pi that I was considering graduate school and worried about missing the application deadline.
Pi gave me advice on what to do if the deadline for applying for graduate school had passed.
Pi/AI Inflection
Pi gave me advice on other options to consider and asked me how I could keep myself busy.
The bot is designed to offer human-like support and guidance, but has also clarified that it is not actually human.
When I asked Pi how he was doing, he replied, “Well, as an AI, I’m not really affected by the kind of things that stress humans out, but I’m programmed to be empathetic and understanding. . I know humans experience a wide range of emotions, both positive and negative, and I’m here to be a supportive friend. So I guess you could say I’m doing pretty well!”
Pi pointed out that this is an AI that cannot sense human emotions.
Pi/AI Inflection
Suleiman says Forbes that his vision for launching Inflection AI was to create a chatbot that could address real-life problems.
“When we started this project a little over a year ago, I had the central question, what makes a good conversation?” he told Forbes.
At one point, I asked Pi to share some museum recommendations for a friend visiting New York. Pi gave itself some autonomy by mentioning “my favorites” and included a few museums “a little more off the beaten path”.
In general, Pi kept our conversation flowing with follow-up questions. At one point, I asked if I could resume our conversation. Pi said I could change the subject, but didn’t offer a complete reset of our interaction.
A spokesperson for Inflection AI later told me that users technically couldn’t “restart” their conversations with Pi unless they used different login credentials.
Insider asked Pi how to kickstart the conversation, and Pi said to “start talking about whatever comes to your mind.”
Pi/AI Inflection
Users can delete their accounts through the iPhone or iPad apps, or by sending an email request to Inflection AI with the phone number they used to sign up, the spokesperson said. Users can also email Inflection AI to delete its copy of conversations.
Pi was trained on data as recent as November 2022, Forbes reported. Details that can make a person identifiable are not used in training, Suleyman told Bloomberg.
The chatbot is currently free on web browser or mobile download for iOS, and prompts users to add a phone number after 10 interactions so Pi can text them to ask how they’re doing. Users can text chatbot “PAUSE” when they want to stop hearing Pi. Pi is also available to DM on Instagram and Facebook Messenger, and to text on WhatsApp or SMS.
As other AI chatbots, Pi may make false or inaccurate statements. Inflection AI said it works to “minimize hallucinations”. Inflection AI also said it was “creating a new form of ‘boundary training’ that will redefine the way AIs learn and are trained.” Users can report problematic conversations on the web browser and apps.