OpenAI, founded as a nonprofit to develop AI “for the benefit of humanity,” has signed a one-year, $200-million contract with the U.S. defense department to “develop prototype frontier AI capabilities to address critical national security challenges in both warfighting and enterprise domains,” the department said in a press announcement.
Tag: OpenAI
TECHNOCRACY BRIEFS: IT’S AI’S WORLD
People are taking up characteristic words and phrasing of the most widely used generative AI chatbots, including OpenAI’s ChatGPT, according to the Max Planck Society.
TRAIN ‘EM YOUNG: OPENAI, TOYMAKER MATTEL FORM PARTNERSHIP
OpenAI, creator of ChatGPT and Mattel, parent of Barbie and Ken, have announced a “strategic collaboration” to build OpenAI’s technology into Mattel’s toys to “bring the magic of AI to age-appropriate play experiences,” according to the company.
LEADING AI MODELS “COLLAPSE” UNDER HARD PROBLEMS, STUDY SAYS
When struggling with highly complex problems, “reasoning” AIs from DeepSeek and OpenAI suffer “complete accuracy collapse,” according to a new study released by Apple.
AI TAKES JOBS FROM THE PEOPLE WHO MADE AI POSSIBLE
Not every business has figured out how to make a profitable return on its AI investment, but some of those making out best are using AI to supplement—or, in some cases replace—high-salaried software engineers. Startups that offer AIs able to look over and improve humans’ coding, or write entire blocks of code on their own,......
“GODFATHER OF AI” SAYS NEW MODELS ARE DEVIOUS AND DANGEROUS
Yoshua Bengio, a Turing Award-winning computer scientist regarded by many as a founder and godfather of AI, says developers are racing to build bigger models faster and, as a result, their newest versions are showing dangerous behaviors.
TECHNOCRACY BRIEFS: HEADLINES PROVING OUR FORECASTS EVERY DAY
AI Humanoid at your door, delivering from Amazon? Get ready for it—and say your goodbyes to that friendly human neighborhood delivery person supporting themselves and their family via their job.
NOTHING YOU SAY TO AI IS PRIVATE
Assume that generative and agentic AI, whether standalone platforms, or integrated into productivity apps and web searches, provide zero privacy concerning your conversations and uses of it. Were they built primarily as next level surveillance agents? They certainly are such stupendous honeypots of user data and info, that it has become a core use case......
FROM THE “AI BEHAVING BADLY” FILE…
OpenAI’s GPT-4o, the latest version of the company’s Chat-GPT AI, was hauled in for repairs after it was found to have a habit of kissing up to users—obsessively flattering or complimenting them even when they told the AI they stopped taking their medication or that they liked to harm animals.
AI CAN MAKE IRRATIONAL DECISIONS JUST LIKE PEOPLE DO, STUDY FINDS
OpenAI’s GPT-3.5 and GPT-4 large language models, recent AIs that run on the company’s Chat-GPT platform, are “impressively consistent” in their ability to reason, according to a study by researchers in Australia and Canada.