![Scientific American Logo](https://cdn.statically.io/img/static.scientificamerican.com/dam/m/3b809e7ff17e4252/original/GettyImages-1926797033_WEB.jpg?w=450)
Please Be Polite to ChatGPT
The benefits of being polite to AI may include prompting better chatbot replies—and nurturing our humanity
Please Be Polite to ChatGPT
The benefits of being polite to AI may include prompting better chatbot replies—and nurturing our humanity
China-U.S. Science Collaborations Are Declining, Slowing Key Research
The U.S. and China are collaborating less on projects across scientific disciplines amid a culture of fear in both countries
Sign up now to get 60 days of digital access
Massive CrowdStrike Tech Outage Highlights Global Vulnerabilities
Companies and governments alike need to step up cybersecurity practices in the wake of massive technology failures associated with a CrowdStrike update
Worldwide Tech Outage Started with Defective Crowdstrike Update to Microsoft Windows
An issue with a commonly used security software called Crowdstrike shuttered large technology systems around the globe, including airlines, transit systems and stock exchanges
Paris Olympics Will Be a Training Ground for AI-Powered Mass Surveillance
In the run-up to the Paris 2024 Olympics, the French government has authorized wide-reaching use of AI software in security surveillance feeds
Blasting Virtual Aliens Could Help Dyslexic Kids Parse Words
Children at risk of dyslexia who played Space Invaders Extreme 2 showed improvement at a word-identification task
Boiling Macaroni in Space? You’ll Need a Weirdly Shaped Pot
Astronauts still survive on freeze-dried meals. Could better food, aided by cooking gadgets designed to be used in microgravity, help them to thrive?
ChatGPT Isn’t ‘Hallucinating’—It’s Bullshitting!
It’s important that we use accurate terminology when discussing how AI chatbots make up information
Can One Chatbot Catch Another’s Lies?
A new approach uses language models to interrogate other language models and sniff out lies
Can AI Be Superhuman? Flaws in Top Gaming Bot Cast Doubt
By learning exploits from adversarial AI, people could defeat a superhuman Go-playing system
Drones Are Doing the Dirty, Dangerous Work of Search and Rescue
As drones get less expensive and computer vision systems improve, rescuers are getting help from artificial eyes in the sky
We Cannot Cede Control of Weapons to Artificial Intelligence
I watched United Nations delegates debate AI-based weapons that can fire without human initiation. Humans cannot be taken out of that decision-making