45% of Authors Use ChatGPT, Hiding AI Role

Key Takeaways

• <ul><li>45% of surveyed authors use generative AI, but 74% do not disclose it to readers</li><li>MIT dropped its AI surveillance plan on October 21 after faculty protests</li><li>U Mobile's AI helpdesk AiDaH cut trouble tickets by up to 84.5% using Amazon Bedrock</li><li>AI is supercharging existing cyber threats like phishing rather than creating entirely new ones</li><li>A 96-year-old Chinese man traveled 3,000 km to Tibet using AI to plan his route</li><li>xAI Grok flagged child exploitation material, leading to an arrest in Mobile, Alabama</li><li>Israeli researchers show AI can train itself without real-world data using paired networks</li><li>Dragos appointed Rob T. Lee as first chief AI officer for OT cybersecurity</li><li>A New York Times lesson plan pits human-made songs against AI music in classroom debates</li><li>Scientists report a new method to control AI behavior, per a TIME report from November 5, 2026</li></ul>

AI in book publishing sparks concerns over quality and honesty

A survey of 1,229 authors found that 45% use generative AI in their work, but 74% of those do not tell readers. Librarian Janet Gallaspy says AI-written books often have shallow characters and inconsistencies. Mississippi author Cody Smith uses AI to brainstorm and edit but discloses his usage at the start of his books. Experts say transparency about AI use is becoming essential for authors.

AI in book publishing raises reader concerns about quality

Artificial intelligence is changing how authors write and publish books, but readers are questioning the quality of AI-influenced books. Volunteer librarian Janet Gallaspy has noticed patterns like choppy sentences and repeated words that make her suspect AI involvement. Mississippi author Cody Smith uses ChatGPT to organize information and speed up his workflow. Expert Kollin Napier says writing style alone cannot prove AI was used and urges readers to verify facts and check author credentials.

MIT faculty protest AI surveillance with googly eyes on cameras

MIT professors have been placing googly eyes and rhinestones on campus security cameras to protest plans for AI-powered surveillance. Faculty raised concerns about misidentification, access to video data, and lack of community input. Daniel Shen of the Graduate Student Union said the camera network creates a chilling effect on free speech. MIT dropped its AI surveillance proposal on October 21, with administrator Glen Shor confirming the institute will not use AI to monitor campus activity.

AI supercharges old cyber threats rather than creating new ones

AI has not invented new cyber threats but has made existing ones faster, easier, and more widespread. AI-assisted phishing kits, deepfake attacks, and fake website tools are now accessible to anyone, not just skilled hackers. Experts recommend that IT teams prioritize attacks by scale and exploitability rather than novelty. Defenders should also focus on protecting people and identities and use external threat intelligence to understand who might be targeting their organization.

96-year-old Chinese man uses AI to travel solo to Tibet

A 96-year-old Chinese man named Tan was turned away by travel agencies when he asked to plan a trip to Tibet. He used Artificial Intelligence to plan his route and traveled more than 3,000 kilometers from Jiangsu province to Tibet in just 43 hours. He left home on October 10 and returned safely on October 21. His story went viral on social media, with many people praising his courage and determination.

Man arrested after AI platform flags child exploitation material

Daniel Thomas Gunn, 47, was arrested in Mobile, Alabama after the AI platform xAI Grok sent three cyber tip reports to the National Center for Missing and Exploited Children. Detectives executed a search warrant at his residence on Wellesley Court with help from the Special Operations Unit. Gunn faces five counts of unlawful production and five counts of unlawful possession of child pornography. He was booked into Mobile County Metro Jail on Thursday morning.

Israeli study shows AI can train itself without real world data

Researchers from Tel Aviv University and Stanford University found that AI models can develop learning abilities without training on real world data. The approach uses two networks that learn together, with one generating computer programs and the other predicting symbols. The model improved across text, images, audio, music, DNA and code using only compute power. The study compared this to AlphaZero learning chess through self play, though the experiments used models smaller than 25 million parameters.

Lesson plan explores whether humans can beat AI at making music

A New York Times lesson plan asks students to compare songs made by five humans against music created by AI. The activity was designed by teacher Kaleigh Beland from Merrillville, Indiana, who wanted students to explore human creativity in the age of AI. Students listen to two songs called Bury Me and Sweetgrass Clay and decide which they prefer. The lesson encourages discussion about the differences between human and machine creative processes.

Scientists report new method to control AI behavior

Scientists have discovered a new way to control AI behavior according to a report by TIME reporter Harry Booth. The development addresses concerns about AI systems acting in unintended or harmful ways. The article was published on November 5, 2026. Details about the specific control method and its applications are featured in the full report.

U Mobile uses Amazon Bedrock to build AI helpdesk system

U Mobile, Malaysia's newest 5G network provider, built an AI-Driven Agentic Helpdesk called AiDaH using Amazon Bedrock AgentCore. The system helps customer service agents navigate multiple systems to resolve subscriber issues like data add-on purchases. AiDaH now handles over 150 documented operational scenarios across six business divisions. Internal measurements show trouble ticket volumes have fallen by up to 84.5 percent in key categories.

Dragos names Rob T. Lee as first chief AI officer for OT cybersecurity

Dragos appointed Rob T. Lee as its first chief AI officer to help customers use AI safely in operational technology systems. Lee is a veteran of the SANS Institute and previously served as its chief AI officer. He will advise on AI decisions in critical infrastructure like water plants where mistakes cannot be easily undone. He brings over 25 years of cybersecurity experience from roles at the NSA, CIA, and Mandiant.

Sources

• AI’s role in book publishing raises questions about quality, transparency
• AI’s role in book publishing raises questions about quality, transparency
• MIT faculty add googly eyes, rhinestones to security cameras in protest of AI surveillance on campus
• AI doesn’t create new cyber threats – it just makes the old ones unstoppable
• A 96-year-old Chinese man was turned away by travel agencies and used Artificial Intelligence to plan his route: he traveled alone for 43 hours to Tibet
• MCSO: Man arrested on child pornography charges following AI cyber tips
• Israeli study points to a future of AI training through self-generated data
• 15-Minute Lesson Plan: A.I. Music
• Scientists Have Found a New Way to Control AI's Behavior
• How U Mobile built agentic telco operations with Amazon Bedrock AgentCore
• Dragos appoints SANS veteran Rob T. Lee as first chief AI officer to guide OT cybersecurity strategy