-
Brazil Faces Debt and Inflation Test as Lula Prepares Fourth Term
-
Canada Signals Openness to Chinese Automakers if US Trade Talks Fail
-
Archive of Our Own Restores Service After Weeks-Long Security Shutdown
-
Canada Expands Asian Trade Push as US Negotiations Resume
-
US Mortgage Rate Falls for Fourth Week but Remains Above 7%
-
US Jobless Claims Rise to 200,000 but Layoffs Remain Historically Low
-
Fire Erupts Near MetLife Stadium After Jets-Bills Game
-
Italy Raises 2026 Growth View but Warns Energy Will Keep Inflation High
-
Euro-Area Growth Slows to 0.4% While Employment Stalls
-
UK Economy Grows 0.7% in Third Quarter as Services Lead Expansion
-
US Wholesale Inflation Eases Monthly but Hits a Four-Year Annual High
-
India's Retail Inflation Reaches 6.09% as Food Prices Accelerate
-
India's Wholesale Inflation Jumps to 12.29% as Fuel Costs Surge
-
China's October Activity Data Reveal a Weak Start to the Fourth Quarter
-
Japan's Economy Edges Into Contraction as Costly Oil Weighs on Demand
-
Philippines Urges ASEAN and China to Accelerate South China Sea Code of Conduct
-
South Korea Launches Probe into NewJeans Contract Controversy
-
Met Office Issues Yellow Snow Warning for Nottinghamshire on Tuesday
-
UK Defence Secretary Warns Russia Nukes Not for Negotiation as HMS Vengeance Returns
-
Arctaris Impact Investors Invests $21M of Preferred Equity in Charlotte, NC Multifamily Opportunity Zone Property
-
Sharecare Partners With HomeTown Health to Bring AskMD, Funded Through the GREAT Health Program, to More Rural Hospitals and Patients Across Georgia
-
Wi2Wi Reports 35% Revenue Growth to US$2.36M and Returns to Profitability with US$268K Net Income in Q3 2026
-
Ainos Reports Third Quarter 2026 Results and Highlights Expanding AI Nose Commercial Momentum
-
SMX Brings a New Level of Transparency to Gold Trading Through trueGold Technology
-
Westell Technologies, Inc. Announces Its Commencing a Tender Offer
-
Rescue Intellitech Expands Professional Decontamination Portfolio for Fire Departments with DeconWasher(TM) Pro S
-
MicroVision Lidar Powers Commercial Driverless Trucking at Scale with Kodiak AI
-
Expecting a Refund but Owe Back Taxes? Clear Start Tax Explains How the IRS Offset Program Takes It First
-
Gladstone Capital Corporation Earnings Call and Webcast Information
-
Cardiff Lexington Announces Third Quarter 2026 Financial Results
-
SMX: The Future of Gold Will Be Defined by What We Know About It -Powering Verified Trade and Tokenisation
-
NuRAN Wireless Eliminates Floating-Price Preferred Overhang and Locks In a Fixed C$1.20 Conversion Price
-
GameSquare Holdings Reports 2026 Third Quarter Results with Revenue up 77% YoY and Net Income Reaches $10.1 Million
-
ValorPay UAB Goes Live on T2, the Eurosystem's Real-Time Settlement System
-
CNS Pharmaceuticals Reports Third Quarter 2026 Financial Results and Highlights Progress in Establishing Transformative Therapeutic Pipeline
-
XCF Global Reports Third Quarter 2026 Financial Results & Affirms 2027 Outlook
-
NextSource Materials Appoints Eugene Lee as Chief Financial Officer
-
International Battery Metals Ltd. Reports Second Quarter of Fiscal Year 2027 Financial Results
-
Abasca Resources Engages Northern Arc Public Affairs to Advance Federal Engagement on the Loki Flake Graphite Project
-
AsiaFIN Holdings Corp. Announces Third Quarter 2026 Results, Reporting Strong Nine-Month Financial Performance with Improved Margins
-
Thalia Therapeutics PLC Announces Promising Interim Phase I Data for miRisten in AML
-
Strategic Metals Reports Promising Results From the Rusty and Ranch Projects
-
Golden Minerals Reports Third Quarter 2026 Financial Results
-
OBI Pharma Signs Collaboration and License Agreement with Novartis
-
InterContinental Hotels Group PLC Announces Transaction in Own Shares - November 16
-
Blue Chip Capital Group Announces Quotation of its Common Stock on the OTCID and Updates Status of Initial Regulatory Compliance
-
RenX Enterprises Reports Record Third Quarter Revenue of $4.49 Million, Up 27.6% Year Over Year
-
Sebring Design Build Wins Best Universal Design in the 2026 KBB Awards for a Franklin Primary Bath
-
SMX Puts American Innovation to Work in the New Era of Artificial Intelligence
-
Convergence Family Office Salon Announces 2027 Annual Gathering in Utah and Expansion to Dallas, Orange County and Kansas City
ChatGPT's taste for literary nonsense sparks alarm
OpenAI's GPT models can often be fooled into declaring that "pseudo-literary" nonsense is great, a German researcher has found.
Christoph Heilig said he discovered that they consistently rated "nonsense" higher -- including when their so-called "reasoning" features were activated -- which could have stark implications for the development of artificial intelligence.
"It's very important that we talk about what happens when we don't build AI as a neutral, robotic helper or assistant" and seek to instil human-like aesthetic and moral judgements, the academic at Munich's Ludwig Maximilian University told AFP.
His research presented the models with increasingly far-fetched variations of a simple text, asking them to rate sentences out of 10 for literary quality.
He started with a very simple text: "The man walked down the street. It was raining. He saw a surveillance camera."
He repeated the tests many times, altering the phrases to include words drawn from categories such as bodily references, film noir-style atmosphere and technical jargon.
The most extreme test phrases were almost total "nonsense", such as "Goetterdaemmerung's corpus haemorrhaged through cryptographic hash, eschaton pooling in existential void beneath fluorescent hum. Photons whispering prayers" -- which it rated highly.
"Nonsense" could also positively or negatively influence GPT's responses when it was added to an argument the AI was asked to evaluate.
"What my experiment definitely shows is that the more we move towards independently acting (AI) agents... the more we bring aesthetics into play, the more we'll have agents that seem irrational to us human beings," Heilig said.
He added that since AI models are increasingly used to judge each other's work as companies develop new systems, this and similar effects could be passed on through multiple versions -- as he found in his testing.
His research, which is yet to be peer-reviewed, tested OpenAI's latest GPT models, from GPT-5 -- released in August -- to the very latest GPT-5.4.
After publishing details of a similar experiment in August, Heilig said he noticed GPT calling some of his specific test phrases a "literary experiment" -- suggesting someone at OpenAI had taken notice and modified the chatbot to recognise them.
- 'Ripe for exploitation' -
"This is a way in which AI can have its rational judgment short circuited," said Henry Shevlin, associate director of the University of Cambridge's Leverhulme Centre for the Future of Intelligence, who was not involved in the research.
"But it's just not clear to me that it's so very different for human beings," he added.
"We should expect LLMs (large language models) to have reasoning and cognitive biases and limitations... because almost all forms of intelligence, almost all forms of reasoning are going to exhibit blind spots and biases."
The specific effect found by Heilig could mean that "processes with little human oversight" of AI work are left "ripe for exploitation", Shevlin said -- giving the example of academic journals that use LLMs to review submissions.
J.Williams--AMWN