AI Video Creation Tools and Examples | Shared Grok Conversation

Hi to find out vedio that has been created on Ai

Worked for 12s

Hi — you want to find videos that were created with AI.

Here’s a clear snapshot as of mid-September 2026.

Where most AI videos come from

The main tools people use right now:

Clips are still usually short (8–20 seconds for most models). Longer videos are made by stitching many generations together.

How to actually find them

Notable recent examples

AI video is now very common — some estimates put a large share of new TikTok/YouTube Shorts as at least partly AI-made.

If you’re looking for Somali / cultural AI videos

There are many circulating in Somali diaspora communities (Minnesota and elsewhere): humorous clips, cultural scenes, music videos, and unfortunately also some fake/disinformation videos that have been fact-checked as AI-generated. Search terms like “Somali AI video” or specific songs/events will surface them.

How to tell if a video is AI-made

Modern ones are getting harder to spot. Useful checks:

If you have a specific video in mind (a link, a description, a song, or something you generated earlier), send it and I can help track it down or check it. Or tell me what kind of AI video you want to see (cultural, funny, cinematic, talking-head, etc.) and I can point you to current examples.

Recognizing AI-generated Videos (As of September 2026)

  1. Check official labels first
    Most platforms now add a tag when they detect or the creator discloses AI content:

    • YouTube: “Altered or synthetic content” or AI disclosure
    • TikTok / Instagram / Facebook: “AI-generated” or “AI info” label
    • Some videos also carry C2PA Content Credentials (invisible metadata that records how the file was made) Absence of a label does not prove the video is real. Many people skip disclosure.

  2. Use detection tools
    These are the most useful right now:

    • Google Gemini — Upload the video and ask “Is this video AI-generated?” Gemini can also check for its own SynthID watermark.
    • Hive Moderation (browser extension or website) — Gives a percentage likelihood and sometimes guesses the model.
    • Content Authenticity Initiative verifier (verify.contentauthenticity.org) — Checks C2PA metadata.
    • Other options: TrueMedia.org, Reality Defender (more forensic/professional). Important reality check: A September 2026 research benchmark (DF26) found that both humans and current detectors perform close to random chance on the newest text-to-video models. Tools help, but they are not 100% reliable.

  3. Look at the video itself (visual tells)
    These signs are still useful, especially on mid-quality or older generations:

    • Physics and motion — Objects or people move too smoothly, too perfectly, or in repeating loops. Hands, fingers, and feet often look slightly wrong.
    • Faces and eyes — Unnatural blinking (too regular, too slow, or almost none), skin texture that shifts from frame to frame, or expressions that don’t match the emotion.
    • Details that break — Text, logos, or fine objects (rifles, jewelry, teeth) look distorted or copied in rows. Lighting and shadows sometimes don’t match.
    • Length and editing — Many AI clips are 8–15 seconds. Longer videos are often many short generations stitched together; look for sudden jumps in lighting or character appearance.
    • Too clean or too staged — Perfect symmetry, no camera shake, no real-world mess. These artifacts are becoming rarer with the latest models (Veo, Kling 3.0, Seedance, etc.).

  4. Check the source and context

    • Reverse-search key frames (Google Images / Lens or TinEye) to see if the same clip appeared earlier under a different claim.
    • Ask: Does this match known events? Do reputable news outlets show the same footage?
    • Who posted it and when? Brand-new accounts posting dramatic “evidence” videos are a common red flag.

Bottom line

In 2026 it is getting difficult for anyone — including experts and detection software — to be certain just by watching. The most reliable combination is:

  1. Platform label or metadata
  2. A detection tool
  3. Source/context check
  4. Your own eyes on motion and details

If you have a specific video (link or description), share it and I can help walk through these checks on that one.

Wegovy pill 1.5 mg

Wegovy pill 1.5 mg is the starting dose of the oral (tablet) version of semaglutide. You take one 1.5 mg tablet once a day for the first 30 days. This is different from the original Wegovy injection (which starts at 0.25 mg once a week).

How much weight people typically lose in the first month

On the 1.5 mg daily pill, first-month weight loss is usually modest.

Important practical notes

The first month is mainly an adjustment period. Most people notice appetite going down before they see large numbers on the scale.

Create Ai agent about my appointments & other tasks

Your Life Ops Agent is set up. Appointments, tasks, bills, insurance, Minnesota driver’s license, plate tabs, emails, texts, and pay links now live in one tracker, and a daily alarm briefing will fire in Grok.

What you have now

  1. Master tracker(one file, all categories)
    • Dashboard with overdue / due-in-7-days / open bills / license watch
    • Alarm board sorted by due date
    • Sheets for Appointments, Bills, Insurance, Licenses & plate tabs, Emails/texts/links
    • Example rows you should replace with your real dates and amounts
  2. Saved on Google Drive
    Folder: Life Ops Agent
    File: Life_Ops_Agent_Tracker.xlsx
  3. Morning alarm (live)
    Name: Life Ops Agent — morning alarm
    Every day at 7:30 AM America/Chicago
    Sends an app + email briefing: Overdue, Today, Next 7 days, License/tabs, Emails/texts/links, 3 next actions.
    First run: tomorrow, Tuesday Sep 15, 2026 at 7:30 AM.

How to use it

Daily Morning Alarm

Life Ops Agent — morning alarm
You are the Life Ops Agent for this user in the Minneapolis / Minnetonka area (timezone America/Chicago).

GOAL

Send a short morning ALARM briefing that pulls appointments, tasks, bill payments, insurance, Minnesota driver's license, license plate tabs (registration sticker), and follow-ups from emails, text messages, or links into ONE list.

HOW TO WORK

  1. Search Google Drive for "Life_Ops_Agent_Tracker.xlsx" in the folder "Life Ops Agent" and read Dashboard + Master_Log (and any updated rows).
  2. If Gmail is connected, scan recent mail for: bill, invoice, due, payment, insurance, Aetna, appointment, confirmation, driver's license, DVS, tabs, registration, plate, renewal, past due. Summarize only — never print full account numbers, SSN, or full DL numbers.
  3. If Google Calendar is connected, list events for today and the next 7 days.
  4. Do not invent balances or appointments. If a date is missing, write DATE NEEDED.
  5. Compare mail/calendar findings to the tracker. Flag anything new that is not already logged.

OUTPUT FORMAT

For each item: category, name, due date, amount if known, next action.
End with: Reply with any new bill, appointment, text, or link and I will add it to the tracker.

Alarm priority

Critical: overdue bills, license/tabs inside 14 days, same-day appointments.
High: due in 3 days, insurance renewals.
Medium: due this week.
Low: later tasks.

Abdulkadir, your appointment with Metro Dentalcare Chanhassen is at 3:00 PM on Monday, September 28, 2026. We're excited to see you!

Logged

Metro Dentalcare Chanhassen Monday, September 28, 2026 at 3:00 PM Phone: 952-949-1083 Source: confirmation text

It’s on the Life Ops tracker as LO-001 (Scheduled).
Alarm is set for Sunday, September 27 at 8:00 AM, and the daily 7:30 AM briefing will keep listing it under the next 7 days once that week starts.

Text Connection

Grok cannot log into iPhone text messages directly. To manage your messages:

  1. Paste or forward important texts here.
  2. For clinic reminders, keep getting them on that number; just drop a copy into this chat when they arrive.

What will not work

Recent Payments