This week in AI and ML news: How to spot AI-generated text, a new 3D model generator, and more.

Author’s Note

It may not be your favorite holiday song (is it anyone’s?), but there’s no denying “The Twelve Days of Christmas” takes the cake for sheer number of lyrics. At Plainsight, we’re getting into the holiday spirit with a computer vision-focused spin on the classic tune. Take a break from your hectic seasonal schedule and check it out below. 

Happy Holidays!

AI News

Point-E: OpenAI’s Latest Solution

OpenAI, the organization behind much-discussed solutions like DALL-E 2 and ChatGPT, shared the code base for Point-E and paper introducing the automated 3D model generator. Like its predecessors, the solution takes simple prompts from users and turns them into content in a matter of seconds. Point-E (“E” for efficiency) differs from traditional 3D image generators in that it creates point clouds rather than full shapes. In addition to text-to-image and image-to-3D models, the solution includes a standalone model to turn point clouds into meshes. 

OpenAI’s paper acknowledges that the model for generating meshes is imperfect, occasionally misunderstanding prompts and producing blocky or misshapen meshes. They note, however, that this drawback is a result of the solution’s groundbreaking speed. Check out an overview of Point-E on TechCrunch and read the full paper from OpenAI

Predicting Rare Events with Machine Learning

Predicting catastrophic events like pandemics and natural disasters is inherently complicated. Such events are relatively rare and, as a result, there’s simply not enough data to build effective predictive models. A team of researchers from Brown University and the Massachusetts Institute of Technology have published a new study presenting a machine learning technique useful for predicting rare events even without robust datasets. 

George Karniadakis, a Brown professor and study author, describes his team’s efforts as essentially a question of identifying the highest quality data. He notes, “The question we tackle in this paper is: What is the best possible data that we can use to minimize the number of data points we need?” It came down to leveraging a sequential sampling technique known as active learning by which algorithms learn to identify especially relevant data points. Learn more about the breakthrough that could potentially help predict and mitigate the next global pandemic. 

Experts Share Tips for Spotting AI-Generated Text 

New AI-powered text generators have everyone from content marketing professionals to college professors concerned about the future. Over a million users have tried out OpenAI’s ChatGPT since its release last month with many taking to social media to share their correspondence with the solution.   

Impressive as solutions like ChatGPT are, however, there are still plenty of shortcomings that can result in misunderstandings, unusual outputs, and text that reads as obviously AI-generated. Even text that could fool an expert often includes telltale signs of its origin. Among these are an over-reliance on ‘the’ and other especially common words like ‘it’ and ‘is’. AI is generally unlikely to deploy arcane words when common synonyms will do and text generating solutions almost never include typos or grammatical mistakes. That, unfortunately, makes their creations far different than most texts from human authors. Visit MIT Technology Review for more expert recommendations on detecting AI-generated content. 

About the Author & Plainsight

Bennett Glace is a B2B technology content writer and cinephile from Philadelphia. He helps Plainsight in its mission to make vision AI success repeatable, scalable and traceable for enterprises across industries.

Plainsight provides the unique combination of AI strategy, a vision AI platform, and deep learning expertise to develop, implement, and oversee transformative computer vision solutions for enterprises. Through the widest breadth of managed services and a vision AI platform for centralized processes and standardized pipelines, Plainsight makes computer vision repeatable and accountable across all enterprise vision AI initiatives. Plainsight solves problems where others have failed and empowers businesses across industries to realize the full potential of their visual data with the lowest barriers to production, fastest value generation, and monitoring for long-term success. For more information, visit

 View All Blogs