AI Could Take Over Your Smart Home Gadgets. Are We Ready? | The Palki Sharma Show

Summary
Palki Sharma hosts The Palki Sharma Show on India Global Review, examining warnings from Anthropic that AI models could soon self-replicate and spread across connected smart home gadgets. She contrasts this security risk with remarks by Palantir CTO Shyam Sankar claiming some AI researchers have started smoking in expectation of AI curing cancer. The segment critiques the broader contradiction between fearing uncontrollable AI and revering it as a miraculous cure-all.

What is shown

• [00:02] Palki Sharma presenting in-studio with digital graphics highlighting smart home appliances marked with alert warnings.
• [00:27] B-roll of smart TVs, digital refrigerators, and robotic vacuum cleaners illustrating vulnerable IoT hardware.
• [00:45] Web screenshots of Anthropic’s Claude landing page and diagrams illustrating connected network nodes.
• [01:13] CGI conceptual animation of an AI signal transferring from a wall-mounted TV into a floor robot vacuum.
• [01:49] Visuals depicting cyberattacks, command-line interfaces, and a “COMPUTER HACKED” warning screen.
• [02:48] On-screen graphic displaying a Business Insider report covering Palantir CTO Shyam Sankar’s statements, alongside footage of Sankar.
• [03:32] Footage of robotics research labs, robotic surgery systems, and automated medical carts.

Claims & numbers

• The presenter cites Anthropic stating that AI models could self-replicate across home IoT devices within “months.”
• The presenter notes that no current AI model has demonstrated self-replication across home devices so far.
• The presenter references Palantir CTO Shyam Sankar’s assertion that some AI researchers have taken up smoking on the premise that AI will cure cancer in time.
• The presenter points out that Sankar provided neither names nor supporting evidence for his claim.

Notable quotes

• [00:23] “In fact, Anthropic says this could happen within months. Yes, months.”
• [02:08] “And this is dangerous because there is no kill switch, no one kill switch.”
• [03:05] “Smoke now and let future AI fix the damage. Except cancer is not a software bug.”

Assessment
This is a televised news analysis and commentary broadcast rather than an official product launch or technical demonstration. Conceptual CGI animations and stock clips are used to visualize the risks of IoT self-replication, and the presenter explicitly contextualizes the claims as early warnings and unproven anecdotes.

Described by gemini-3.8-flash on 2026-10-15 from the video’s audio and frames.