Diplomacy & Crisis News

After the initial near-euphoria about Large Language Models, or LLMs, that power generative artificial intelligence (AI), the mood has gone sour. The spotlight shines now on doomsday scenarios where LLMs become self-aware, go out of control, and extinguish humanity.

Fear of sentient robots is hardly new. In an 1899 short story, Ambrose Bierce conjured a robot created by an inventor named Moxon. It looked like a person, if a dour one, but it wasn’t smart enough even to beat Moxon at chess. And when it lost, the robot revealed deep wells of uncontrolled emotion: it murdered Moxon.

This fear has maintained its popularity ever since in books, plays, and movies. Some bad robots appeared simply as machine systems, like homicidal HAL in Stanley Kubrick’s classic 2001: A Space Odyssey. Some robots look human, like the Terminators. And beyond the murderous robots, there are sometimes big networks of robotic systems, such as in The Matrix, whose aim is to enslave humanity. Even Isaac Asimov, who tried to rein in robots with three laws that forbade doing harm to humans, worried that robots could circumvent such strictures.

ChatGPT and Bard are two prominent examples of LLMs that amaze with sophisticated answers to questions. These systems have sparked a huge wave of investment in new services powered by LLMs. And they have unleashed a torrent of anxiety about how their proneness to “hallucinate” (make stuff up) might create havoc with fake news, stolen elections, massive job losses, undermined trust in business, or even destabilization of national security. The worst fears concern the potential for the machines to become sentient and subjugate or exterminate us. A chorus of leading voices from the worlds of high tech and politics has made a case, best summed up by Henry Kissinger, that current advances in AI have put the world in a “mad race for some catastrophe.”

Our assessment is that the furor over the extinction prophecy has gotten the better of us and is distracting from the important work of learning how to use an extremely valuable but inherently error-prone technology safely.

The core of ChatGPT is a huge artificial neural network of 96 layers and 175 billion parameters, trained on hundreds of gigabytes of text from the Internet. When presented with a query (prompt), it responds with a list of the most probable next words. A post-processor chooses one of the words according to their listed probabilities. That word is appended to the prompt and the cycle repeated. What emerges is a fluent string of words that are statistically associated with the prompt.

These strings of words are drawn from multiple text documents in the training set, but the strings do not appear in any single document. ChatGPT is incapable of verifying whether a response is truthful. Its responses that make no sense are called “hallucinations” when all they are is statistical inference from the training data.

Despite their unreliability, LLMs can be useful for amusement and for initial drafts of documents, speeches, research projects, and code. The smart thing is to use them for these purposes but not in any application where harm can result from invalid answers. In fact, it is not hard to imagine harnessing the machine impartiality of ChatGPT to solve contentious problems. For example, we think a robotic approach to gerrymandering would be a great way to build confidence in AI. Task competing LLMs with designing congressional districts that look like simple geometric forms rather than exotic reptiles. The main guidance would be that the districts would have to be as balanced as possible between the registered voters of the two major parties. Our bet is that bots will succeed wildly where humans have failed.

What about the fears of sentience? Can LLMs eventually absorb so much text that they possess all human knowledge and are smarter than any of us? Are they the end of history? The answer is a clear no. The claim that all human knowledge can eventually be captured into machines makes no sense. We can only put into machines knowledge that can be represented by strings of bits. Performance skills like sports, music, master carpentry, or creative writing are prime examples of knowledge that cannot be precisely described and recorded; descriptions of skill do not confer a capability to perform. Even if it could be represented, performance skill is in forms that are inaccessible for recording—our thoughts and reflections, our neuronal memory states, and our neuro-muscular chemical patterns. The sheer volume of all such nonrecorded—and unrecordable—information goes well beyond what might be possible to store in a machine database. Whatever functions can be performed by LLMs are small compared to human capabilities.

In addition to this, statistical inference is surely not the whole story of human cooperation, creativity, coordination, and competition. Have we become so mesmerized by Large Language Models that we do not see the rest of what we do in language? We build relationships. We take care of each other. We recognize and navigate our moods. We build and exercise power. We make commitments and follow through with them. We build organizations and societies. We create traditions and histories. We take responsibility for actions. We build trust. We cultivate wisdom. We love. We imagine what has never been imagined before. We smell the flowers and celebrate with our loved ones. None of these is statistical. There is a great chasm between the capabilities of LLMs and those of human beings.

And beyond LLMs, there is no sign on the horizon of a more advanced, close to intelligent, technology.

So, let’s take a sober attitude toward LLMs, starting by curbing the sensational talk. What if we use the phrase “statistical model of language” instead of “Large Language Model”? Notice how much less threatening, even silly, the extinction prophecy sounds when expressed as, “Humanity goes extinct because of its inability to control statistical models of language.”

Tamping down unreasonable fears will allow us to attend to the serious matters of the economic and social impacts of the latest advances in artificial intelligence, and of LLMs’ penchant for inaccuracy and unreliability. Let us also address the geopolitical stresses between the United States, China, and Russia, which could be exacerbated by an unbridled military arms race in AI that might make going to war seem more thinkable—and which would actually heighten the risks of nuclear escalation by the side losing a machine-based conflict. In this respect, we concur with Kissinger that advanced AI could catalyze a human catastrophe.

Above all, as with previous periods that featured major technological advances, the challenge now is to chart a wise path around fear and hype.

John Arquilla and Peter Denning are distinguished professors at the U.S. Naval Postgraduate School. John Arquilla’s latest book is Bitskrieg: The New Challenge of Cyberwarfare (Polity, 2021). Peter Denning most recently co-authored Computational Thinking (MIT Press, 2019).

The views expressed in this article are solely theirs.

Image: Shutterstock.

U.S. environmental measures have China as an unspoken target.

Without a playbook to turn to, officials are still grasping for solutions.

Anti-nuclear weapons activists have a new best friend: the Catholic Church.

Defense deals and tech ties underpin Modi’s visit to Washington.

Il n'y aura sans doute bientôt plus de terrains vagues dans les grandes villes, et pas davantage de lieux désaffectés, refuges des enfants, des errants sans toit, des artistes à la recherche d'un local même ouvert à tous les vents : l'« urbanisme transitoire » se charge de les transformer en sites (...) / France, Art, Logement, Société, Urbanisme, Ville, Banlieue - 2018/04

Assassinat d'un journaliste trop curieux en Slovaquie, nouveau premier ministre tchèque et chef du Parlement roumain poursuivis pour détournement de fonds : l'Europe centrale et orientale semble accablée par la corruption. Pourtant, pots-de-vin et trafics d'influence ne seraient pas moindres (...) / Pologne, République tchèque, Slovaquie, Criminalité financière, Extrême droite, Nationalisme, Europe centrale, Corruption - 2018/04

Washington must take difficult steps to prevent catastrophe.

A new European recovery program.

The AI Act instills greater privacy standards, stricter transparency laws, and steeper fines for failing to cooperate.

As the Republican presidential primary intensifies, a burgeoning contingent of right-leaning foreign policy experts has emerged to claim that President Joe Biden’s Ukraine policy is eroding America’s ability to deter a Chinese invasion of Taiwan. Consequently, they recommend drastically reducing, if not outright halting, American support for Ukraine’s defense. Politically, this argument is shrewd—it appeals to an increasingly Ukraine-skeptical Republican primary electorate without compromising its proponents’ credibility within mainstream foreign policy circles. Given the widespread consensus that China poses America’s most significant strategic challenge, framing a rollback of current Ukraine policy this way lends the argument an air of hard truth told by sober-minded adults. However, while certainly politically savvy, geopolitically, this line of reasoning is highly unsound.

At its core, the case for reducing U.S. support for Ukraine is based on a supposed policy tradeoff: every dollar or bullet sent to Ukraine is one less for Taiwan’s defense. Because Taiwan’s security is more strategically significant to Washington than Ukraine’s, critics claim the U.S. must realign its policy to match its priorities.

However, this argument overlooks several factors that challenge its fundamental assumptions. Notably, one doesn’t need to accept the view expressed by the Taiwanese, among others, that the United States must support Ukraine to deter China. Even setting aside concerns about U.S. credibility or resolve, there is ample reason to conclude that the critics of Washington’s current Ukraine policy are mistaken. Similarly, while the most compelling argument for assisting Ukraine is arguably the moral one, even within the framework of tradeoffs, the case for the Biden administration’s current policy is strong.

First, the defense budget does not solely consist of spending on Ukraine and Taiwan. Many other programs, some arguably wasteful, could be reduced or eliminated to increase funding for Taiwan’s defense. Furthermore, Russia is China’s most militarily capable partner and would likely be willing to supply weapons to China during any conflict over Taiwan. Therefore, providing Ukraine with the means to destroy Russian military capabilities is possibly the most cost-effective Defense Department program currently in existence, even when one looks at it through the lens of a Taiwan contingency. And while far from guaranteed, if Russia’s poor performance in Ukraine eventually results in Putin’s overthrow and replacement by a more benign Russian government, the gains to U.S. and Taiwanese security would be even more significant.

The radically different nature of the two conflicts further undermines the notion of a sharp tradeoff between arming Ukraine versus Taiwan. The Ukraine conflict is predominantly a land war fought over short distances, while a Taiwan conflict would likely be a primarily naval war fought over long distances. While reports of the depletion of American weapons reserves are worrying, the weapon systems in question are not the submarines, aircraft, anti-ship missiles, sea mines, and torpedoes crucial to Taiwan’s defense.

The U.S. military has also gained valuable information thanks to its support for Ukraine. The war has exposed inefficiencies and gaps in America’s defense industrial base—lessons better learned now than in the middle of a potential conflict with China over Taiwan. The Ukraine mission is also giving the Pentagon a chance to test new equipment and develop the logistical skills necessary for long-distance supply operations, knowledge that would prove invaluable should the United States need to come to Taiwan’s defense.

One might counter that the U.S. commitment to Ukraine will tempt China to strike Taiwan while the United States is tied down in Europe, especially if the conflict drags on for years. But attacking Taiwan while the Ukraine conflict is ongoing would be a strategic disaster for China. Beijing hopes to drive a wedge between Europe and the United States over Taiwan, and as evidenced by recent comments from leaders like French president Emmanuel Macron, they have had some success. However, a Chinese attack on Taiwan amid Russia’s invasion of Ukraine would torpedo that diplomatic effort and likely result in the two conflicts merging into one global East-West struggle. Under such circumstances, Europe would likely support U.S. efforts to defend Taiwan in return for continued American assistance in Ukraine. In short, attacking Taiwan in the context of the Ukraine war makes the emergence of a pan-Western coalition to counter China more likely, not less.

Lastly, there is an essential distinction between the current conflict in Ukraine and a hypothetical one over Taiwan that makes arguments about potential tradeoffs moot: the role of nuclear weapons. The U.S. has wisely refrained from direct involvement in the Ukraine war to avoid the possibility of a nuclear confrontation with Russia. But a serious conflict over Taiwan would necessitate significant U.S. military involvement, likely involving missile strikes on the Chinese mainland. It is illogical to conclude that Washington is correct to worry that direct military engagement in Ukraine could result in a nuclear war with Russia, but killing significant numbers of Chinese citizens does not also risk a nuclear exchange. Misconceptions about the role of atomic weapons in a Taiwan conflict have likely been reinforced by DC think tanks’ war games, which often take them off the table. But in any real-world conflict over Taiwan, the prospect of nuclear war would immediately loom over the minds of policymakers in both Washington and Beijing.

To be clear, the threat of nuclear war with China is not a reason to abandon Taiwan, just as during the Cold War, the threat of nuclear war was neither a reason to abandon Berlin. Nevertheless, the fact that hypothetical mushroom clouds loom over any U.S.-China conflict constrains the role of conventional weapons in any Taiwan scenario.

Those who worry about tradeoffs for Ukraine make the mistake of believing that the United States must exceed China’s conventional capabilities in the Western Pacific to achieve deterrence. Under this logic, it is understandable why one would be desperate to shift every weapon possible from Ukraine to Taiwan. But attempting to match China’s conventional forces this way is untenable and unnecessary.

To deter a Chinese invasion, the United States simply needs to maintain sufficient military capabilities in the region such that China cannot successfully invade Taiwan without simultaneously attacking America’s Pacific bases, particularly Guam, but conceivably Japan and the Philippines as well. Under these conditions, China will find itself confronting a sort of “Guam trigger.” Given the stated U.S. policy, Beijing must operate under the assumption that Washington would actively assist in Taiwan’s defense. If China launches an invasion without first destroying America’s military assets in the region, its ships will be left vulnerable to attack. However, if it launches a preemptive strike on U.S. forces, especially on American soil in Guam, it will experience the full wrath of a vengeful United States. Given this, China faces a choice between a failed invasion or a major conflict likely ending in atomic annihilation. Given those options, Beijing will presumably choose to maintain the status quo, achieving Washington’s goal of deterrence with limited deployment of conventional arms. China may attempt to escape this bind by adopting more restrained tactics, such as a blockade of Taiwan. But any U.S. attempt to counter such a move would primarily involve diplomatic outreach and air and sea lift, areas that are not hampered by U.S. support for Ukraine.

It is undoubtedly correct that U.S. national security officials must recognize tradeoffs, but it is equally true that they must also recognize false tradeoffs when they emerge. The idea that Washington must choose between defending Taiwan and defending Ukraine is one such false choice, and the Biden administration is wise to ignore such criticisms.

Robert Nelson is a Ph.D. candidate at Stanford University studying U.S. foreign policy. He previously served as a national security aide to Senator Chris Murphy. His Twitter handle is @RW_Nelson.

Image: Shutterstock.

Beijing is now the world’s largest holder of international port assets. Through a network of dozens of state-owned port operators, contractors, investment firms, and banks, the Chinese state has invested upwards of $110 billion in foreign port operation and development projects across eighty port states, a value equivalent to the total outward foreign direct investment (ODFI) stock of Israel.

Figure 1. Locations of port investments by Chinese state-owned enterprises.

Figure 2. Chinese state-owned enterprises investing in foreign ports.

These investments were mostly made in the wake of the 2008 Global Financial Crisis, when demand for capital was high and the rehabilitative and mutually beneficial effects of foreign direct investment (FDI) were presumed.

However, the emergence of several troubled Chinese state-led port investments has diminished this optimism. Contentious contract renegotiations, threatened project cancellations, or outright bans on Chinese FDI signal growing unease about “handing over the keys” of one’s critical infrastructure to a foreign government. Such concerns are not unfounded. In a recently published article in the journal Marine Policy (available open access) my co-authors and I argue that the political nature of these investments vis-à-vis Chinese state ownership of the investing firms carries a unique array of economic, strategic, and political risks for the recipient port states.

First, Chinese state-led investments are uniquely susceptible to commercial failure. Overcapacity is a systemic problem in China’s industrial economy, driven by national subsidies and poor internal controls on state-led industrial development. Since 2013, Beijing’s policy on correcting overcapacity has been to offshore it to developing economies through large-scale overseas infrastructure development projects. The result has been several “white elephant”-type port development projects, driven by China’s need to relieve pressure on bloated domestic industries rather than projected commercial benefits for the recipient port state. The most well-known of these is Hambantota Port, whose annual losses of $60 million meant that Sri Lankan authorities were only able to repay the $1.5 billion in Chinese loans used to finance its development by awarding a ninety-nine-year lease to a Chinese state-owned port operator.

Second, Chinese state-led investments risk the host state becoming entrapped in China’s military disputes. As Chinese state control over foreign port assets has grown, Chinese military strategists have increasingly been positioning these assets within Beijing’s naval doctrine. Specifically, these strategists argue that foreign port assets, despite being procured under commercial pretenses, may be appropriated for military missions, ranging from general reconnaissance to serving as logistics and replenishment hubs for Chinese naval vessels. If such military uses were attempted, the host port state would be in the invidious position of either accepting Chinese military operations in its territorial waters and risking retaliation from China’s military rivals, or rejecting Chinese military operations in its territorial waters and risking retaliation from China.

Third, port states hosting Chinese state-led investments risk becoming targets for espionage. Ports are rich targets in this case, both in military and commercial terms. Militarily, espionage can reveal the movement of military hardware through the port as well as the operational characteristics of foreign or domestic navies using the port; for example, resupply and materiel requirements, personnel identities, and origins and onward destinations. Commercially, espionage can provide commercial intelligence on traded consignments and port services which can be mined to provide a competitive advantage to competing firms or identify chokepoints in the industrial strategies of port states. Chinese state actors may, working through Chinese state-owned firms controlling foreign port assets, embed surveillance equipment and personnel into those assets to pursue these military and commercial dividends.

And fourth, Chinese state-led port investments may become vehicles for economic coercion. In times of political conflict between Beijing and the port state, Beijing may instruct its state-owned port investors operating in that state to disrupt port operations, such as by diverting port traffic, halting terminal operations, withholding follow-on funding, initiating vexatious litigation, or terminating the contract. This creates pressure on port states to avoid political conflicts with Beijing to maintain the economic well-being of its Chinese-controlled ports, potentially limiting the port state’s autonomy when dealing with issues sensitive to Beijing. For example, observers have credited Greece’s 2017 veto of a European Union statement on human rights in China on China’s state-owned port operator COSCO taking a controlling stake in Greece’s busiest port the year prior.

While China’s port investments should be encouraged to the extent that they help close the international infrastructure gap, increase legitimate competition among international shipping and logistics providers, and further integrate China into the global economy, port states must remain attentive to the multifaceted risks posed by China’s unique state-led OFDI program.

Dr. Christopher J. Watterson is a Research Fellow at the Department of War Studies, King’s College London.

Image: Shutterstock.

The cargo shows how the global south is balancing relations with Moscow and the West.

The defenses that made the country unique are falling one by one—leaving political discontent and spiritual voids exposed to al Qaeda.

In great-power competition there is no such thing as minding one’s own business.

It’s in Washington’s interest to make the best possible use of Moscow’s barbaric folly.

The weapons may bolster the very rebel groups West African governments are trying to defeat.

A muddled U.S. strategy is making it harder for Kyiv to win.

The country’s ongoing war isn’t a reason not to bring it into the West’s security alliance.

Global equality and its discontents.