It’s time for Bits & Bytes…

… where we bring you news, innovations, and thought-provoking insights from AI, IT, and beyond. In this week’s newsletter we’re looking at:

• Microsoft 365 Adds ‘Deep Research’ AI Tools
• GPT-4o Gets Image Creation & Editing
• Google Launches Gemini 2.5

IN THE SPOTLIGHT

🔍 Microsoft 365 Adds ‘Deep Research’ AI Tools

Microsoft introduces AI-powered research capabilities to its 365 Copilot platform with two new tools: Researcher and Analyst. Researcher leverages OpenAI’s deep research model with enhanced orchestration to handle complex business analyses, while Analyst uses OpenAI’s o3-mini reasoning model to tackle data-intensive problems with Python execution capabilities.

What sets Microsoft’s approach apart is these tools’ ability to access workplace data across connected platforms like Salesforce and Confluence alongside web information, creating a more comprehensive research experience, though challenges remain in ensuring factual accuracy and preventing hallucinations.

TL;DR

• Two new reasoning tools: Researcher and Analyst.
• Access workplace data across connected platforms.
• Can develop strategies and create reports.
• Available to Frontier program members in April.

TECH HEADLINES FROM ACROSS THE WEB

🔮 GPT-4o Gets Image Powers

OpenAI upgrades ChatGPT with native image creation and editing capabilities powered by GPT-4o, a significant improvement over its previous DALL-E 3 system. The feature, which allows for more detailed images and editing of photos containing people, is initially available to $200/month Pro subscribers with plans to roll out to Plus and free users soon.

⚡ Google Launches Gemini 2.5: Answer to OpenAI

Google launches Gemini 2.5 Pro Experimental, a reasoning-focused AI model that pauses to verify facts and think through problems. Available immediately to developers and Gemini Advanced subscribers, the model features an enormous 1-million token context window (expanding to 2 million soon) and outperforms competitors on specific code editing benchmarks.

📸 Apple to Use Maps Imagery for AI Training

Apple quietly updated its privacy policy to reveal plans for using Apple Maps Look Around imagery to train AI models starting March 2025. The company will leverage its vast collection of street-level photos and 3D data to enhance image recognition, creation, and enhancement technologies, though it promises to only use imagery with faces and license plates already blurred.

TECH FOR GOOD

💎 AI Reveals Mineral Treasures Where Humans Missed Them

Earth AI has discovered promising deposits of copper, cobalt, gold, silver, and other critical minerals in previously ignored regions of Australia by using artificial intelligence to analyze decades of archived geological data. This breakthrough demonstrates how AI can revolutionize resource exploration in environmentally responsible ways while addressing the growing demand for materials essential to clean energy technologies.

Interested in contributing a story to next week’s tech newsletter? Hit us up and let’s collab 💥

The post Google Launches Gemini 2.5 – Google’s reasoning-focused AI model appeared first on ChannelBytes.