I remember an SF short story where somebody tried to time travel to the past and ended up way out in space.

[Goonch]

20. You may quote my posts in context if you also quote entirely my following inquiry to and responce from Google AI

AI Speaks For It Self

Question by Goonch to Google AI: Is AI currently dangerous? What is the frequency of misinformation or falty answers/advice? What should people do when AI is a danger to them and how will they know the danger exists?

Answer by Google AI:
"AI currently poses several real-world dangers, primarily through the industrial-scale generation of misinformation, sophisticated deepfake scams, and potential for biased or dangerous advice in high-stakes fields like healthcare

While "sentient" AI remains a hypothetical long-term risk, immediate harms like identity theft and the erosion of digital trust are active threats as of early 2026.

Frequency of Misinformation and Faulty Advice:
The reliability of AI varies significantly by task, but inaccuracies are common:

Fact-Checking Failures: In August 2025, leading AI tools repeated false information about current news roughly 35% of the time, a significant increase from the previous year.

General Inaccuracy: Recent studies indicate that roughly 23% of all AI responses contain some form of inaccurate information, while 31% of automated decisions require human correction.

Hallucinations: Specialized models for grounded tasks like summarization have lower "hallucination" (fabrication) rates, ranging from 0.7% to 4.1% depending on the specific model and complexity of the query.

Medical & High-Stakes Errors: AI has been found susceptible to fabricated data in over 30% of medical prompts in some studies, often due to "hallucinating" citations or misapplying general logic to specific patient cases.

How to Identify AI Dangers
You can spot AI-generated risks by looking for technical and logical "tells":

Visual Glitches: Look for unnatural lighting, blurring around the edges of faces, or inconsistent shadows (e.g., shadows pointing in multiple directions despite a single light source).

Audio Anomalies: Listen for a flat, robotic tone, lack of natural breathing, or unnatural pacing and "slurred" words
.
Logical Impossibilities: In breaking news, check for seasonal mismatches (e.g., leafless trees in a video supposedly from summer) or the presence of anachronistic technology.


Manufactured Urgency: Scammers often use AI to impersonate loved ones or authorities in "emergencies." Be skeptical of any request that demands immediate financial or personal information.

What to Do When AI is a Danger
If you suspect AI is providing dangerous advice or being used to target you:

Demand Context: For high-stakes advice, ask the AI: "Ask me 5 questions about my situation before you give me any advice". If it gives confident answers without needing your specific details, it is likely guessing.

Cross-Verify: Never rely on a single AI output for health, legal, or financial decisions. Use a second opinion from a human professional or trusted, non-AI sources.

Use Verification Tools: Upload suspicious media to platforms like TrueMedia.org to check for deepfake markers, or use tools like TinEye to see if an image is an old one being reused out of context.

Report Harm: If you encounter harmful content or impersonation scams, report it directly to the platform (e.g., social media or the AI provider) and, if it involves identity theft, local law enforcement. "

Please let me know when you post. Thanks, Goonch