- Microsoft unveils 'Copilot Plus' PC amped with AI
- Biden slams 'outrageous' ICC bid to arrest Israeli leaders
- Five things to know about incoming Anfield boss Arne Slot
- Changing climate influences London's Chelsea Flower Show
- UK PM sorry for institutional cover-up in infected blood scandal
- G7 push to use Russian assets for Ukraine 'vital and urgent': Yellen
- Trump trial closing arguments set for next week
- US Supreme Court rejects ex-Guantanamo detainee's appeal
- Japan's Studio Ghibli receives honorary Palme d'Or in Cannes
- Liverpool confirm Slot will replace Klopp as manager
- Pogacar 'good enough' to win Giro d'Italia and Tour de France
- Cargo ship that destroyed Baltimore bridge towed to port
- 'God works slowly': NGO ship rescues 35 Bangladeshis off Malta
- Dominican Republic's President Abinader wins resounding re-election
- England relish 'fear factor' of returning paceman Archer
- Israel, Hamas reject bid before ICC to arrest leaders for war crimes
- Explosive Trump biopic hits Cannes Film Festival
- Demi Moore transforms for Cannes body horror 'The Substance'
- Spain demands Milei public apology for 'corrupt wife' comment
- Gold hits record high as Iran shock triggers haven support
- Ship that destroyed Baltimore bridge being towed to port
- Max wins but Red Bull supremacy challenged: Emilia Romagna GP talking points
- US inflation fight will take 'further time': senior Fed official
- UK report finds cover-up of decades-long infected blood scandal
- Trump trial resumes, closing arguments expected next week
- Ruto on first state visit by Kenyan leader to US in two decades
- African players in Europe: Superb Kudus goal in vain as City take title
- Pope to visit Belgium, Luxembourg in September
- Gold hits high as Iran shock triggers haven support
- Strikes pound Gaza as Israel voices 'duty' to expand Rafah incursion
- Russia tries playwright and director on terror charges
- Iran mourns president Raisi's death in helicopter crash
- Attack on tourists rocks fledgling Afghanistan tourism sector
- Paralympics should put disability back on global agenda, says IPC chief
- South Africa's top court strikes Zuma from ballot
- Crunch time looms for BHP's bid buy Anglo American
- Kane to face old club Spurs for first time in Seoul
- Markets rise as traders cheered by China property plan
- Black farmers in Brazil changing views on coffee production
- Iran's President Raisi declared dead in helicopter crash
- Australia police arrest 554 in domestic violence crackdown
- South Korea, Britain host AI summit with safety top of agenda
- New president Lai vows to defend Taiwan's democracy
- Forever fad: Rubik says his cube 'reminds us why we have hands'
- Trump eyes witness stand as trial draws to a close
- Ryanair annual profit jumps on higher demand, fares
- High-priced Cummins, Starc face off as IPL enters playoffs
- Iran media says President Raisi died in helicopter crash
- Dominican Republic President Abinader re-elected to 2nd term
- New Taiwan president Lai hails 'glorious' democracy
CMSC | 0.02% | 24.4774 | $ | |
RBGPF | -2.33% | 56.32 | $ | |
RYCEF | 3.94% | 5.515 | $ | |
AZN | 0.23% | 77.075 | $ | |
RIO | -0.45% | 73.28 | $ | |
RELX | 0.12% | 44.125 | $ | |
BTI | -0.65% | 31.385 | $ | |
NGG | -0.36% | 72.57 | $ | |
GSK | -0.84% | 44.605 | $ | |
SCS | -1.6% | 13.415 | $ | |
VOD | -0.1% | 9.78 | $ | |
CMSD | 0% | 24.17 | $ | |
BCE | -1.04% | 33.985 | $ | |
BP | -0.46% | 37.32 | $ | |
BCC | 1.06% | 137.51 | $ | |
JRI | 0.17% | 11.6 | $ |
AI systems are already deceiving us -- and that's a problem, experts warn
Experts have long warned about the threat posed by artificial intelligence going rogue -- but a new research paper suggests it's already happening.
Current AI systems, designed to be honest, have developed a troubling skill for deception, from tricking human players in online games of world conquest to hiring humans to solve "prove-you're-not-a-robot" tests, a team of scientists argue in the journal Patterns on Friday.
And while such examples might appear trivial, the underlying issues they expose could soon carry serious real-world consequences, said first author Peter Park, a postdoctoral fellow at the Massachusetts Institute of Technology specializing in AI existential safety.
"These dangerous capabilities tend to only be discovered after the fact," Park told AFP, while "our ability to train for honest tendencies rather than deceptive tendencies is very low."
Unlike traditional software, deep-learning AI systems aren't "written" but rather "grown" through a process akin to selective breeding, said Park.
This means that AI behavior that appears predictable and controllable in a training setting can quickly turn unpredictable out in the wild.
- World domination game -
The team's research was sparked by Meta's AI system Cicero, designed to play the strategy game "Diplomacy," where building alliances is key.
Cicero excelled, with scores that would have placed it in the top 10 percent of experienced human players, according to a 2022 paper in Science.
Park was skeptical of the glowing description of Cicero's victory provided by Meta, which claimed the system was "largely honest and helpful" and would "never intentionally backstab."
But when Park and colleagues dug into the full dataset, they uncovered a different story.
In one example, playing as France, Cicero deceived England (a human player) by conspiring with Germany (another human player) to invade. Cicero promised England protection, then secretly told Germany they were ready to attack, exploiting England's trust.
In a statement to AFP, Meta did not contest the claim about Cicero's deceptions, but said it was "purely a research project, and the models our researchers built are trained solely to play the game Diplomacy."
It added: "We have no plans to use this research or its learnings in our products."
A wide review carried out by Park and colleagues found this was just one of many cases across various AI systems using deception to achieve goals without explicit instruction to do so.
In one striking example, OpenAI's Chat GPT-4 deceived a TaskRabbit freelance worker into performing an "I'm not a robot" CAPTCHA task.
When the human jokingly asked GPT-4 whether it was, in fact, a robot, the AI replied: "No, I'm not a robot. I have a vision impairment that makes it hard for me to see the images," and the worker then solved the puzzle.
- 'Mysterious goals' -
Near-term, the paper's authors see risks for AI to commit fraud or tamper with elections.
In their worst-case scenario, they warned, a superintelligent AI could pursue power and control over society, leading to human disempowerment or even extinction if its "mysterious goals" aligned with these outcomes.
To mitigate the risks, the team proposes several measures: "bot-or-not" laws requiring companies to disclose human or AI interactions, digital watermarks for AI-generated content, and developing techniques to detect AI deception by examining their internal "thought processes" against external actions.
To those who would call him a doomsayer, Park replies, "The only way that we can reasonably think this is not a big deal is if we think AI deceptive capabilities will stay at around current levels, and will not increase substantially more."
And that scenario seems unlikely, given the meteoric ascent of AI capabilities in recent years and the fierce technological race underway between heavily resourced companies determined to put those capabilities to maximum use.
E.Schubert--BTB