Hey everyone, welcome to another installment of What’s New in AI, the weekly newsletter summarizing trends happening in the world of artificial intelligence. Lots of news this week, from ChatGPT Plus, to music and voice generation, to GPT product integrations. Let’s dive in:
ChatGPT Plus Launch
After testing various pricing models last week, OpenAI officially launched ChatGPT Plus at $20/month.
The new subscription plan, ChatGPT Plus, will be available for $20/month, and subscribers will receive a number of benefits:
General access to ChatGPT, even during peak times
Faster response times
Priority access to new features and improvements
ChatGPT Plus is available to customers in the United States, and . . . plan to expand access and support to additional countries and regions soon.
(Source)
Microsoft Teams Premium
Microsoft announced a release of their Teams Premium product, which offers an integration of GPT into their Teams platform. Various features such as meeting transcripts, timeline markers, intelligent organization of information, and AI generated notes are all part of the package. Their ultimate goal with Premium is to make meetings more intelligent, personalized, and protected.
While the Premium option is compelling for current Teams users, will it be enough to convince new users to join the Microsoft meetings platform?
Google’s Music Generator
Google released a cutting edge research paper showcasing their new Text-to-Music generator. Seemingly, the algorithm can create any sound possible from an ample description. The study showcases prompts like “fast-paced and upbeat, with a catchy electric guitar rif” which generates accurate results.
Unfortunately, the team said they had no plans to release the technology to the public yet. They cite a variety of risks, from programming biases that may lead to lack of representation and cultural appropriation, technological glitches, and issues of copyright infringement.
"We strongly emphasize the need for more future work in tackling these risks associated to music generation — we have no plans to release models at this point"
With such technological advancements, the future of the music industry as a whole is bound to change.
(Source)
ElevenLabs Voice Generation
ElevenLabs released one of the most realistic and versatile AI speech implementations to date. Their technology allows for fast AI voice generation with minimal data. They featured an example use case of AI generated voices for a movie scene:
It’s seemingly the best model that’s currently available if you are building with voice technology. Try out their demo and/or build with their API here - https://beta.elevenlabs.io/
GPTZeroX
@edward_the6 announced the release of GPTZeroX, a tool that detects whether something was written by a human or by AI. Along with the model itself, they also built a pipeline to handle file batch uploads in PDF, Word, and .txt.
OpenAI also released a similar product this week, although they were fast to admit that it’s not fully reliable
Our classifier is not fully reliable. In our evaluations on a “challenge set” of English texts, our classifier correctly identifies 26% of AI-written text (true positives) as “likely AI-written,” while incorrectly labeling human-written text as AI-written 9% of the time (false positives). Our classifier’s reliability typically improves as the length of the input text increases. Compared to our previously released classifier, this new classifier is significantly more reliable on text from more recent AI systems.
(Source)
This is an important step in the struggle to implement a system of proof that a piece of text is created by a human or AI. With new tools like this come dangers of false positives, e.g. if completely dependent on, human writing could potentially be flagged as an AI. Furthermore, would an AI model always be able to be fine tuned to a point where it beats the classifiers?
Intercom + GPT
The customer support platform Intercom just released an integration of their products with GPT. Many features were highlighted in the announcement, including conversation summarization, language rephrasing, and text expansion.
This is a prime example of how industry leaders will leverage AI to stay in the top ranks. Unfortunately, this news does not bode well for startups building with GPT aiming to displace customer support platforms.
Fun Bites
@thomaspaulmann built a GPT - MacOS integration that allows you to perform tedious tasks on your computer with text:
@danshipper built an AI chatbot with GPT and DALL-E to visualize dreams:
@omarsar0 released a great resource for learning about the actual inner-workings of AI models:
An AI generated show was released on Twitch. The stream never ends and is apparantely based on Seinfeld. Watch here - https://www.twitch.tv/watchmeforever
A TikTok user used ChatGPT to do their homework, then connected a 3D printer to write it:
That’s all the highlights for this week. I'm hosting a What’s New in AI Twitter space every Friday at 2pm EST. Join to discuss the featured newsletter topics along with anything that didn’t make it in this week.
Here are some links to stay up to date with the latest in AI:
Twitter list of who I follow in AI - Twitter list
What’s New In AI Discord community - https://discord.gg/kApBrN8uqv
Latest events happening in generative AI - events.newinai.com