AI could soon pass test that once proved humans are smarter | Latest Tech News
This system could sport us.
Artificial intelligence is already outperforming humans at varied intelligence-based actions ranging from chess to sample recognition. Now, specialists declare they’re a yr away from beating “Humanity’s Last Exam” (HLE) — a supposedly unsolvable test that only our best and brightest can pass.
“Model builders have really done a great job at improving these reasoning models,” Calvin Zhang, the research lead at Scale, the AI firm behind HLE, told The Times of London.
“Humanity’s Last Exam stands as one of the clearest assessments of the gap between AI and human intelligence,” declared Dr. Tung Nguyen, a pc science and engineering professor at Texas A&M who contributed 73 of the questions (the second most). Mojahid Mottakin – stock.adobe.com
Developed to see how close AI is to the “frontiers of human expertise,” this intelligence benchmark is comprised of 2,500 questions spanning over 100 extremely specialised fields, ranging from mythology to rocket science.
Over 1,000 authorities from across the sciences, humanities and arts contributed to the HLE, which was designed to required PHD-levels of comprehension to ace — just past the experience of AI, Nueroscience News reported.
Zhang said the final word objective was to create a “closed-ended academic benchmark, set to the frontier of expert humans, that only a handful of people on Earth can really solve.”
Nonetheless, AI’s efficiency on the HLE has improved at exponential speeds within a short period of time. While ChatGPT answered fewer than 3% of questions accurately during its first attempt in 2024, its rival Google Gemini obtained 18.8% of the questions proper within months.
Last month, that quantity improved to over 45%.
Anti-AI activist teams led a ”March Against The Machines” through King’s Cross in the UK to advocate for a global pause on the development of superior artificial intelligence. ZUMAPRESS.com
Zhang believes that AI could strategy full marks — anybody scoring close to 100% is outlined as a “universal expert” within a yr.
“If we truly cared about this as the only thing in life, I think we could get to it pretty quickly,” boasted Kate Olszewska, a product supervisor at Google DeepMind.
Kate Olszewska, a product supervisor at Google DeepMind, agrees: “If we truly cared about this as the only thing in life, I think we could get to it pretty quickly.”
This light-speed progress is spectacular given the pains Scale took to make the HLE AI-proof. The test-makers reportedly supplied a $500,000 prize to specialists who could contribute questions that could not be simply answered via web search, finally drawing over 70,000 responses.
They made sure the questions couldn’t be answered via a simple online search. Ascannio – stock.adobe.com
Any questions that could be answered by current fashions have been discarded until the examination was whittled down to 2,500 of the most AI-ironclad queries. For occasion, testees is perhaps requested to translate historical Palmyrene inscriptions or to determine microanatomical buildings in birds during the course of the test examination,
To additional make sure the test was AI-ironclad, the group saved most of the solutions hidden so that later fashions couldn’t memorize them.
“Humanity’s Last Exam stands as one of the clearest assessments of the gap between AI and human intelligence,” declared Dr. Tung Nguyen, a pc science and engineering professor at Texas A&M who contributed 73 of the questions (the second most).
He argued that while some of the aforementioned fashions carried out properly, the poor scores of the remaining illustrate that the chasms between AI and human intelligence stay “wide.”
“When AI systems start performing extremely well on human benchmarks, it’s tempting to think they’re approaching human‑level understanding,” Nguyen said. “But HLE reminds us that intelligence isn’t just about pattern recognition — it’s about depth, context and specialized expertise.”
The techspert said that the final word objective wasn’t to stump “AI,” but to relatively to illustrate the systems’ strengths and weaknesses.
In flip, this would help us construct “safer, more reliable technologies” while also demonstrating “why human expertise still matters” — an important objective in a world where AI appears to be changing us in every sector from fast food to medication.
That being said, AI has displayed a surprisingly humanlike aptitude for downside fixing, demonstrating that its processing powers aren’t relegate to rote reminiscence.
In 2025, assessments by Chinese researchers revealed similarities between the AI fashions’ “perception” and human cognition — notably when it got here to language grouping.
From this, researchers deduced that the machine learners “develop human-like conceptual representations of objects.”
“Further analysis showed strong alignment between model embeddings and neural activity patterns” in the area of the mind related with reminiscence and scene recognition.
Stay informed with the latest in tech! Our web site is your trusted source for breakthroughs in artificial intelligence, gadget launches, software program updates, cybersecurity, and digital innovation.
For contemporary insights, professional coverage, and trending tech updates, go to us repeatedly by clicking right here.
