AI News Drop by CHI Software

?? Audible's AI Voice Training

Audible is using AI trained on professional narrators to generate new audiobook recordings. Select narrators will train the AI with their voices to approve and edit their synthetic voice for specific works. These performers will be compensated for audiobooks created using their AI voices.

?? Microsoft Tackles Deepfakes

Microsoft has partnered with StopNCII to help remove synthetic nude images from Bing search results. This tool uses a digital fingerprint to identify and scrub explicit images from the platform. Microsoft has already taken action on 268,000 images in a pilot project.

?? First Impressions of OpenAI o1

OpenAI released its new o1 models, designed to "think" before answering complex questions. o1 excels at reasoning but is more expensive than GPT-4o and needs some tools and speed. The model is best for complex tasks with a step-by-step reasoning approach.

?? Apple's Approach to AI

Apple Intelligence aims to offer more efficient device performance and make your daily tasks faster. So, it’ll include rephrasing text, summarizing emails, and generating clip art. The most important points are convenience and privacy in AI usage.

?? Meta’s AI Info Label Changes

Meta is changing how it labels AI-edited content on Instagram, Facebook, and Threads. The "AI info" label will move to the post's menu for edited content but remain under fully AI-generated content. This change aims to reflect better the extent of AI used in content.

?? Saudi Arabia's AI Processing Center

Groq Inc. and Aramco are building the world's largest AI processing center in Saudi Arabia with 19,000 language processors, which may expand to 200,000 units. It's a plan to reduce the country’s dependence on oil and bring AI development in the Middle East, Africa, and India.

Do you have any news to add? Feel free to share them in the comments!