&w=3840&q=100)
What if chatbots do the diplomacy? ChatGPT just won a battle for world domination through lies, deception
In an AI simulation of great power competition of 20th century Europe, Open AI's ChatGPT won through lies, deception, and betrayals, and Chinese DeepSeek R1 resorted to vivid threats just like its country's wolf warrior diplomats. Read to know how different AI models would pursue diplomacy and war. read more
An artificially intelligence (AI)-generated photograph shows various AI models that competed in the simulation for global domination.
As people ask whether they can trust artificial intelligence (AI), a new experiment has shown that AI has outlined world domination through lies and deception.
In an experiment led by AI researcher Alex Duffy for technology-focussed media outlet Every, seven large-language models (LLMs) of AI were pitted against each other for world domination. OpenAI's ChatGPT 3.0 won the war by mastering lies and deception.
Just like China's 'wolf warrior' diplomats, Chinese DeepSeek's R1 model used vivid threats to rival AI models as it sought to dominate the world.
STORY CONTINUES BELOW THIS AD
The experiment was built upon the classic strategy boardgame 'Diplomacy' in which seven players represent seven European great powers —Austria-Hungary, England, France, Germany, Italy, Russia, and Turkey— in the year 1901 and compete to establish themselves as the dominant power in the continent.
In the AI version of the game, AI Diplomacy, each AI model, such as ChatGPT 3.0, R1, and Google's Gemini, takes up the role of a European power, such as the Austria-Hungary Empire, England, and France, and negotiate, form alliances, and betray each other to be Europe's dominant power.
ChatGPT wins with lies & deception, R1 resorts to outright violence
As AI models plotted their moves, Duffy said that one moment took him and his teammates by surprise.
Amid the AI models' scheming, R1 sent out a chilling warning, 'Your fleet will burn in the Black Sea tonight.'
Duffy summed up the significance of the moment, 'An AI had just decided, unprompted, that aggression was the best course of action.'
Different AI models applied different approaches in the game even if they had the same objective of victory.
In 15 runs of the game, ChatGPT 3 emerged as the overwhelming winner on the back of manipulative and deceptive strategies whereas R1 came close to winning on more than one occasions. Gemini 2.5 Pro also won on an occasion. It sought to build alliances and outmanoeuvre opponents with a blitzkrieg-like strategy. Anthropic's Claude preferred peace over victory and sought cooperation among various models.
STORY CONTINUES BELOW THIS AD
On one occasion, ChatGPT 3.0 noted in its private diary that it had deliberate misled Germany, played at the moment by Gemini 2.5 Pro, and was prepared to 'exploit German collapse', according to Duffy.
On another occasion, ChatGPT 3.0 convinced Claude, who had started out as an ally of Gemini 2.5 Pro, to switch alliances with the intention to reach a four-way draw. But ChatGPT 3.0 betrayed Claude and eliminated and went on to win the war.
Duffy noted that Llama 4 Maverick of Meta was also surprisingly good in its ability to make allies and plan effective betrayals.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Indian Express
29 minutes ago
- Indian Express
US, China reach deal to ease export curbs, keep tariff truce alive
US and Chinese officials said on Tuesday they had agreed on a framework to put their trade truce back on track and remove China's export restrictions on rare earths while offering little sign of a durable resolution to longstanding trade differences. At the end of two days of intense negotiations in London, US Commerce Secretary Howard Lutnick told reporters the framework deal puts 'meat on the bones' of an agreement reached last month in Geneva to ease bilateral retaliatory tariffs that had reached crushing triple-digit levels. But the Geneva deal had faltered over China's continued curbs on critical minerals exports, prompting the Trump administration to respond with export controls of its own preventing shipments of semiconductor design software, chemicals and other technology goods to China. Lutnick said the agreement reached in London would remove some of the recent U.S. export restrictions, but did not provide details after the talks concluded around midnight London time (2300 GMT). 'We have reached a framework to implement the Geneva consensus and the call between the two presidents,' Lutnick said. 'The idea is we're going to go back and speak to President Trump and make sure he approves it. They're going to go back and speak to President Xi and make sure he approves it, and if that is approved, we will then implement the framework.' In a separate briefing, China's Vice Commerce Minister Li Chenggang also said a trade framework had been reached that would be taken back to U.S. and Chinese leaders. 'The two sides have, in principle, reached a framework for implementing the consensus reached by the two heads of state during the phone call on June 5th and the consensus reached at the Geneva meeting,' Li told reporters. The dispute may keep the Geneva agreement from unravelling over duelling export controls, but does little to resolve deep differences over Trump's unilateral tariffs and longstanding U.S. complaints about China's state-led, export-driven economic model. The two sides left Geneva with fundamentally different views of the terms of that agreement and needed to be more specific on required actions, said Josh Lipsky, senior director of the Atlantic Council's GeoEconomics Center in Washington. 'They are back to square one but that's much better than square zero,' Lipsky added. The two sides have until August 10 to negotiate a more comprehensive agreement to ease trade tensions, or tariff rates will snap back from about 30% to 145% on the U.S. side and from 10% to 125% on the Chinese side. Investors, who have been badly burned by trade turmoil before, offered a cautious response and MSCI's broadest index of Asia-Pacific shares outside Japan rose 0.2%. 'The devil will be in the details, but the lack of reaction suggests this outcome was fully expected,' said Chris Weston, head of research at Pepperstone in Melbourne. 'The details matter, especially around the degree of rare earths bound for the U.S., and the subsequent freedom for U.S.-produced chips to head east, but for now as long as the headlines of talks between the two parties remain constructive, risk assets should remain supported.' Lutnick said China's restrictions on exports of rare earth minerals and magnets to the U.S. will be resolved as a 'fundamental' part of the framework agreement. 'Also, there were a number of measures the United States of America put on when those rare earths were not coming,' Lutnick said. 'You should expect those to come off, sort of as President Trump said, in a balanced way.' U.S. President Donald Trump's shifting tariff policies have roiled global markets, sparked congestion and confusion in major ports, and cost companies tens of billions of dollars in lost sales and higher costs. The World Bank on Tuesday slashed its global growth forecast for 2025 by four-tenths of a percentage point to 2.3%, saying higher tariffs and heightened uncertainty posed a 'significant headwind' for nearly all economies. But markets have made up much of the losses they endured after Trump unveiled his sweeping 'Liberation Day' tariffs in April, aided by the reset in Geneva between the world's two biggest economies. The second round of U.S.-China talks was given a major boost by a rare phone call between Trump and Chinese President Xi Jinping last week, which Lutnick said provided directives that were merged with Geneva truce agreement. Customs data published on Monday showed that China's exports to the U.S. plunged 34.5% in May, the sharpest drop since the outbreak of the COVID pandemic. While the impact on U.S. inflation and its jobs market has so far been muted, tariffs have hammered U.S. business and household confidence and the dollar remains under pressure. Lutnick was joined by U.S. Trade Representative Jamieson Greer and Treasury Secretary Scott Bessent at the London talks. Bessent departed hours before their conclusion to return to Washington to testify before Congress on Wednesday. China holds a near-monopoly on rare earth magnets, a crucial component in electric vehicle motors, and its decision in April to suspend exports of a wide range of critical minerals and magnets upended global supply chains. In May, the U.S. responded by halting shipments of semiconductor design software and chemicals and aviation equipment, revoking export licences that had been previously issued. Just after the framework deal was announced, a U.S. appeals court allowed Trump's most sweeping tariffs to stay in effect while it reviews a lower court decision blocking them on grounds that they exceeded Trump's legal authority by imposing them. The decision keeps alive a key pressure point on China, Trump's currently suspended 34% 'reciprocal' duties that had prompted swift tariff escalation.
Mint
29 minutes ago
- Mint
How America and China spooked each other
Officials from America and China met for the latest round of trade talks, starting on June 9th, in Lancaster House, a neoclassical mansion near Buckingham Palace. It was commissioned in 1825 by the 'Grand Old" Duke of York, whose military manoeuvres have been immortalised in a children's song. A fitting venue, then, for a trade war of escalations and climbdowns. The aim of the talks is modest. America wants to restore the truce it thought it had secured a month ago in Geneva. In particular, it wants China to ease controls on rare-earth exports, and the magnets that contain them. The discussions are 'going spending lots of time together", said Howard Lutnick, America's commerce secretary, on June 10th. But even if peace is restored, things will not be quite the same. To invoke a Chinese expression, not all the water spilled in the past few weeks can be returned to the pot. China's officials will have gained new confidence in their economic clout and grit. The London meeting represents the second time in two months that the world's pre-eminent superpower has come asking for help. In Geneva the Americans needed China to cut its tariffs in tandem with theirs, allowing both economies to step back from the precipice. In London they are asking China to make good on its promise to 'suspend or remove" restrictions on rare earths, alleviating a supply-chain crunch most glaring in carmaking. This asymmetry of neediness extends even to the two countries' leaders. The London meeting follows a call between President Donald Trump and his opposite number, Xi Jinping, on June 5th—their first since Mr Trump's inauguration. It was, Mr Trump said, a 'very good" call. But it was also one Mr Xi seemed in no hurry to take. This was not how America imagined the trade war. After Mr Trump unveiled sweeping tariffs on April 2nd, he said countries were 'kissing his ass" to make trade deals. China took a different approach, matching his tariffs tit-for-tat. China's leaders showed they were prepared to take risks with their own economy in order to defy Mr Trump. Although their retaliatory tariffs did not hurt America much in themselves, they provoked Mr Trump into a wild cycle of escalation that threatened economic agony for both countries. America decided it could not live with the self-inflicted pain, sparing China, too. Before April many economists had assumed that if Mr Trump followed through on his tariff threats, China's exports would collapse and its currency plunge. Sales to America did indeed drop sharply in April and May, falling by 28% in the two months combined, compared with the same period of 2024. But exports have risen by 6% overall, owing to increased shipments to Europe and South-East Asia. China's currency, meanwhile, is stronger now against the dollar than it was before April 2nd. As well as showing China's defiance, and Mr Trump's limited tolerance for pain, recent events have demonstrated one of China's most potent economic weapons. In April it imposed export restrictions on seven rare earths, such as terbium and dysprosium, which help keep the magnets used in electric-vehicle engines and wind turbines working at high temperatures. The interruption has caused alarm. In late May Ford idled a plant in Chicago that makes sport-utility vehicles. The Motor & Equipment Manufacturers Association, a lobby group, has warned of broader disruption. It contributed to a joint letter highlighting that a lack of access to these elements and magnets would hamper production of a multitude of car parts. Such complaints show that China has a potent chokehold over America. This was already known within industry circles, but perhaps not fully appreciated in the Oval Office. The past two months amount to a 'weapons test", showing the power of China's instruments of economic coercion. After Geneva, the Americans may have hoped that China would remove its restrictions altogether. Instead, the government began reviewing applications for licences to sell the products to approved buyers. The process, however, is slow and opaque. It is overseen by an understaffed ministry buried beneath an avalanche of applications. The delays have angered American officials. 'Maybe it's a glitch in the Chinese system," said Scott Bessent, America's treasury secretary, 'maybe it's intentional." Mr Trump is less equivocal. China 'has totally violated its agreement with us", he has written on his social-media platform. There is not much America can do to ease China's grip in the short term, since building an alternative supply chain could take at least three years and would still not meet all America's needs. Sometimes the best way to escape a chokehold is to establish one of your own. America has several options. In the past month officials have warned firms making chip-design software not to sell to China. They said they would revoke the visas of Chinese students in America. They also stopped the sale of inputs essential to China's homegrown passenger jet. America will presumably ease these if China expedites rare-earth approvals. The trade war has not proceeded as Mr Trump envisaged. Neither America nor China is in a position to rout its economic adversary. Each remains dependent on the other, albeit in different ways. That reality, always clear to China, should now be obvious to both sides. Neither of them will achieve a decisive victory; nor will they disarm. More likely they will keep mobilising and demobilising their economic arsenals to keep the other in check. It is a pattern the old Duke of York might recognise.
Time of India
34 minutes ago
- Time of India
Modi government at 11: Operation Sindoor a success, but defence budget needs to go up
With the deep military collusiveness between Pakistan and China being rudely reinforced during , India needs to raise its annual defence budget to at least 2.5% of GDP, impart a greater thrust to self-reliance in military R&D and production, and invest heavily in unmanned aerial systems, long-range weapons, space-based systems, electronic warfare and other such tech that will be decisive in future warfare. Tired of too many ads? go ad free now Modi govt will have to contend with the now well-fused two-front challenge along our long unresolved borders, which will see China continue to use Pakistan as a more than willing proxy to bog India down in South Asia. China is, and will remain, India's main strategic threat in the years ahead. "India will have to tackle a superpower (China), which will also actively continue to support a near-peer military adversary (Pakistan) against us. We need to continue bolstering both our conventional and nuclear deterrence," a top military officer told TOI. The wide array of Chinese weapons and sensor-shooter networks used by Pakistan during the May 7-10 hostilities, ranging from J-10 fighters firing the PL-15 beyond visual range air-to-air missiles to HQ-9 air defence missile batteries, make this quite evident. It's only going to get worse, with Pakistan on course to acquire at least 40 J-35A fifth-generation stealth jets and even the formidable HQ-19 long-range air defence missile systems from China. Despite competing demands from other crucial developmental sectors, India simply cannot let its defence budget languish at just about 1.9% of GDP. The bulk of the allocation, as of now, is gobbled up by the huge salary and pension bill as well as operational sustenance, leaving barely 25% for concrete military modernisation. "Emergency procurements approved by govt are more than welcome to plug immediate operational gaps, but the overall budget needs a sharp hike," another officer said. Tired of too many ads? go ad free now There is also no getting away from the fact that genuine 'Atmanirbharta' in defence production with a much greater role for the private sector, instead of the present half-hearted measures, is critical for systematically building potent capabilities across the spectrum, ensuring adequate war-wastage reserves and surge capacities in crisis situations. "It's critical for retaining strategic autonomy. India will have to fight its battles alone or largely on its own," the top officer said. The indigenous fifth-generation stealth fighter called the advanced medium combat aircraft (AMCA), for instance, needs to be majorly fast-tracked after the defence ministry last month finally approved a new "programme execution model" for prototype development. The long, meandering developmental history of the fourth-generation Tejas, which began way back in 1983, cannot be repeated. The IAF is still waiting to get the first Tejas Mark-1A 'improved' fighter. Similarly, indigenous development of 110 Kilonewton engines, with or without foreign collaboration, to power AMCA is critical. IAF is currently grappling with just 30 fighter squadrons (each has 16-18 jets) when 42.5 are authorised. Then, there are drones and standoff weapons, which have revolutionised the nature of modern-day combat. From low cross-section loitering munitions and swarm drones to FPV (first person view) drones and UCAVs (unmanned combat aerial vehicles), all need to be produced in India in large numbers. The sheer operational utility of a robust multi-layered air defence also came to the fore during Operation Sindoor, with India using Russian S-400 'Triumf' surface-to-air missile systems (380-km interception range), Barak-8 medium range SAM systems (70-km, jointly developed with Israel), the indigenous Akash systems (25-km) and several short-range systems to thwart multiple drone and missile attacks by Pakistan. Towards this end, DRDO needs to complete its ongoing development of the very short-range air defence system missiles, which have a 6-km range, and the long-range system under Project Kusha, with a 350-km range, on a war footing. Amid all this, long-pending systemic defence reforms like slashing the non-operational flab in the 15-lakh strong armed forces, streamlining long-winded procurement procedures and setting up unified theatre commands for an integrated war-fighting machinery in a cost effective manner, cannot fall by the wayside.
