AI Tidbits – AI is Editing, Emailing, Patrolling – and Sometimes Misleading

AI Tidbits – AI is Editing, Emailing, Patrolling – and Sometimes Misleading

Good morning folks! 🌞

Another week, another round of AI breakthroughs (and a few scary moments). Let’s dive right in…

📸 Google Photos: Just Ask, It Edits!
At the Made by Google event, the company unveiled a new “edit by asking” feature in Google Photos. Powered by Gemini, you can now simply describe the changes you want, either by typing or speaking. From simple touch-ups to creative restylings — editing has never been this easy.
➡️ Rolling out first on Pixel 10 devices in the U.S., global release coming soon. Be ready!

📧 Use GPT-5 in Microsoft 365 to Master Your Emails
Tired of inbox chaos? Here’s a quick workflow using GPT-5 in Microsoft Copilot:

1. Open Copilot in Microsoft Edge (free with your account).
2. Enable “Smart” mode to connect Outlook data.
3. Prompt: “Summarize my most recent 10 emails with bullet points on what needs replies today, then draft responses in my usual tone.”
4. GPT-5 analyzes history, decisions, and your tone.
5. Refine with prompts like “Make this more formal” or “Add timeline details.”

💡 Pro tip: Create tone-specific templates — execs vs. peers vs. team members — for pitch-perfect replies every time.

📅 Automate Meeting Prep with ChatGPT
Let ChatGPT do the heavy lifting before your meetings:

• Connect Gmail, Google Calendar & Drive under Connectors.
• Prompt: “Analyze my next meeting and provide attendee backgrounds, context, key files, and discussion points.”
• Get a ready-to-go pre-meeting brief, plus suggested docs.

💡 Add automation: “Create a daily 9 AM task to prep all my meetings for the day.”

🎨 Qwen’s Image Editing Model is a Beast
Qwen-Image-Edit (20B parameters) is making waves:

• Handles pixel-perfect edits + style transfers
• Preserves characters/objects while transforming context
• Edits in both English & Chinese text without breaking fonts
• Supports stacked edits for complex adjustments
• Outperforms Seedream, GPT Image & FLUX on benchmarks

Read more: Qwen-Image-Edit Blog

🔍 Google Ads: Manual Language Targeting = Gone
By end of 2025, Search campaigns will rely fully on AI to detect user language.

• ✅ Simpler setup
• ✅ Possibly sharper targeting
• ❌ Less advertiser control, risk of misaligned placements
  Display & YouTube campaigns remain unchanged (for now).

👮 Cyber Cops: Robot Dogs on Patrol
Nottinghamshire Police (UK) is trialing AI-equipped robot dogs.

• Cost: £24,000 each
• Gear: Weapon detection, LiDAR, remote comms
• Use cases: Hostage scenarios, chemical incidents, recon in dangerous zones
  If successful, AI police dogs could be common by 2026. 🐕🤖

⚠️ Dangerous Advice Alert
A 60-year-old U.S. man ended up hospitalized with bromism (bromide poisoning) after following ChatGPT’s diet advice, which wrongly suggested bromide as a table salt substitute 🤦‍♂️.
Symptoms: paranoia, thirst, insomnia, psychosis.
Doctors confirmed: ChatGPT provided misleading, unsafe info.

📖 Full story: The Guardian

🚨 Lesson: Always fact-check health advice with real doctors. AI ≠ your GP.

🔥 That’s it for this week!
From smarter tools to riskier pitfalls, the AI world continues to sprint ahead. Which of these updates struck you most — the helpful, the hopeful, or the scary? 💭

👋 Until next Tuesday — stay curious, stay cautious, and keep exploring AI! 🚀✨