India has just notified strict new rules for AI and deepfakes. Specifically, the government wants to stop the spread of fake news and audio. Therefore, social media firms must now act much faster to remove harmful posts. These changes will officially start on February 20, 2026.

What are the New AI Rules?

First, the rules define “Synthetically Generated Information” or SGI for the first time. Specifically, this covers any audio or video made or changed by AI tools. Then, these items must look real or true to fall under the law. Actually, routine editing like color fixes or spell checks is not included. Therefore, the law targets content that could trick a person into believing it is real.

The 3-Hour Takedown Mandate

First, platforms must remove flagged AI content within three hours of an order. Specifically, this applies when the government or a court sends a notice. Then, this is a huge drop from the old 36-hour time limit. Actually, platforms must also act on user complaints within a much shorter time now. Therefore, firms like Meta and X need to deploy faster systems to stay legal.

Mandatory Labels and Metadata

First, all AI-made content must carry a clear and visible label. Specifically, you cannot hide these markers in small text or deep menus. Then, platforms must embed permanent metadata into every AI file. Actually, this helps trace where the fake content first came from in the digital world. Therefore, removing these labels or metadata is now a crime for the platforms.

User Declarations and Verification

First, you must declare if your post is AI-made before you upload it. Specifically, apps will now show a new pop-up box for every single post. Then, the platforms will use AI tools to check if you are telling the truth. Actually, firms can be held liable if they let unlabelled AI content go viral. Therefore, users who lie about their AI posts may face a ban or police action.

IT Rules Comparison – 2021 vs 2026

Reality Check

Reality Check: A three-hour window is very hard for any human team to manage. Still, the government says it is the only way to stop viral fake news. In fact, most of this work will be done by automated AI bots. Therefore, some real posts might be taken down by mistake in the rush. Recently, many experts have warned that these rules could limit free speech online. Now, the tech giants must find a way to balance speed with accuracy.

The Loopholes

The Loopholes: The rules do not apply to “good faith” research or teaching materials. Actually, you can still use AI for art if it does not look like a real person. In fact, satirical or parody content may have slightly different rules to follow. Therefore, clear jokes or cartoons are usually safer from these strict removal laws.

What This Means for You

Recently, the rise of deepfakes has caused many cases of online fraud and hate. Now, you should double-check any video that looks too strange to be true. First, look for the new “AI-Generated” label on your social media feed. Then, always declare your own AI edits when you upload a video or photo. Next, report any harmful deepfakes you see to the platform right away. Indeed, these new rules are a big step toward making the internet a safer place.

Next Steps

Check your Instagram or X app for the new AI declaration tool this week. Then, read our guide on how to spot a deepfake video using simple tips. Would you like me to find the full list of “unlawful acts” mentioned in the new IT rules?

LATEST :- 

• EPFO Interest Credit Explained: Earn ₹52,000 on Your PF Balance
• Zero Tax Income: How to Pay Nothing on ₹14.66 Lakh in 2026
• How to Apply for Aadhaar for a Child Below 5 Years
• Draft Income-tax Rules 2026: New Forms and Simplified Filing

The post India’s New AI Rules 2026: 3-Hour Deadline for Deepfake Removal first appeared on Rightsofemployees.com.