I’ll keep you up-to-date with the latest tools and updates. There’s a lot to cover …
The UK’s AI Safety Institute releases Inspect, an open source toolset designed to simplify the safety assessment of AI models. Inspect can be used to test the capabilities of AI models, such as core knowledge and reasoning.
Krea AI now offers its paying users the ability to create videos using AI. The new feature combines both still images and text input.
Enkrypt has published a ranking of the most secure large language models (LLMs) to help companies choose the most suitable models. OpenAI’s GPT-4-Turbo tops the list with the lowest risk score, while models such as Saul Instruct-V1 and Phi3-Mini-4K are at the bottom of the list.
Canadian startup Upend offers a new kind of AI search engine that gives users access to over 100 different language models to get accurate answers with sources. Unlike competitors such as Perplexity, Upend relies on a pricing model designed to appeal to teams and students.
Meta’s new AI tools allow advertisers to create not just backgrounds, but entire images to generate product variations. However, this feature also carries the risk of tricking customers with non-existent products, although Meta emphasizes that it has taken strict security precautions.
Amazon aims to simplify generative AI development with Bedrock Studio: The new tool is designed to help businesses experiment, share, and build AI-powered applications.
Atlassian Rovo is a new AI-powered tool that acts as a knowledge base for organizations by collecting data from different apps and helping employees find information, learn from it, and take action. Rovo is designed to increase productivity by simplifying the search for information within an organization and enabling automated tasks through AI agents.
Nvidia’s ChatRTX chatbot now supports Google’s Gemma model, allowing users to interact with their own documents, photos, and YouTube videos. The update also includes voice search and offers more ways to search locally stored data using different AI models.
One day before Google’s I/O, OpenAI tried to steal the show from its big competitor. And their demos definitely caused quite a stir. The focus was on their latest AI model GPT-4o, where the “o” stands for “omnimodel”. This is to indicate that this version does not only process text, but also e.g. image and …
As expected, Google used the keynote at its I/O developer conference to demonstrate its strength in AI. Among other things, the company presented new AI models for a wide range of tasks. Some will run directly on Android devices or can be found in the Chrome browser. Others use Google’s specialized servers. They create text, …