Rogue (pronounced rohg)
(1) As
“rogue state” or “rogue regime”, of or noting the government of a nation that
ignores or defies international law, treaties and conventions of behaviour.
(2) A
dishonest, unprincipled, untrustworthy person; a rascal; a scoundrel (as a
label, applied overwhelmingly to men).
(3) A
playfully mischievous person; a scamp (sometimes as “loveable rogue”, a
designation sometimes a bit generous).
(4) In
Australia, a horse, mule, or donkey that is difficult to control; a refractory
horse, especially a racehorse (described as a “barrier rogue”, explained by the
beast being “difficult” about being put into the starting gates prior to the
race starting).
(5) A tramp
or vagabond (archaic); it was used also as a verb in the sense of “to live or
act as a scoundrel, vagabond or scamp.”
(6) In
computing, “rogue software” is malware deceitfully represented as anti-spyware.
(7) In role-playing
games, a character class focusing on stealthy conduct. Characters said to be roguelike possess some
but not all the elements of rogues (rogue-lite, either a synonym of roguelike
or an even less roguesque). Roguelikes usually
have permadeath and randomized levels. Permadeth is seemingly (and hopefully)
gaming-specific and describes either (1) a death that compels a player restart
the game from the beginning or (2) a situation where a player character or
characters who die are lost permanently but do not trigger a “game over”,
allowing the player to proceed without the character(s). In a nice touch, this is triggered usually
only if deaths are “avoidable” (ie not forced by the plot), occurring only as a
result of player-error.
(8) In biology,
a usually inferior organism (most used in botany of plants varying markedly
from the normal). In that context, “to
rogue” or the act of “rouging” is to uproot or destroy those plants diseased or
not conforming to the desired standard; this may be of individual plants or one
could “rogue” a whole field, paddock etc.
In industrial horticulture, there are now AI (artificial intelligence)
machines able autonomously “to rouge”, killing weeds or “rogue plants” using
lasers.
(9) By
extension, any inferior or defective specimen.
(10) As “to
go rouge”, “going rogue”, “gone rogue” etc, to stop being obedient or
predictable; to behave in a way that is not allowed or expected.
(11) No
longer obedient, belonging or accepted and hence not controllable or
answerable; a renegade.
(12) Dangerous
and unpredictable (of people, animals, plants & objects). Specifically of animals, having an abnormally
violent or unpredictable disposition (applied often to older males that have separated
from their herd or community to lead a solitary, aggressive life).
(13) To
cheat (archaic).
1555–1565:
Of uncertain origin, the theories include it being from the earlier English roger (a begging vagabond who pretends
to be a poor scholar from Oxford or Cambridge (thus the idea of the “idle
vagrant” or “sturdy beggar”; those who prefer to beg though easily capable of
work)) and that may have been an adaptation (as agent noun) from the Latin rogāre (to ask), the idea being “rogues”
were “fake beggars” who affected poverty to gain sympathy and coins from those
passing. That seems plausible (Latin a
popular source for such coinings) although in Latin “to beg or demand”
something from someone was petere while
specifically to beg or entreat for money was pecūniam ōrāre. One who
wandered from town-to-town as a beggar or trickster was an aeruscāre and, reflecting prejudices still seen today, the word was
also applied to fortune-tellers and itinerate buskers. An alternative etymology is as a direct
re-purposing of the Middle French rogue (arrogant,
haughty), from the Old Northern French rogre
(aggressive), from the Old Norse hrokr
(excess, exuberance), said to be inherited as the Icelandic hroki (arrogance) although not all
lexicographers agree. The Celtic use is
attributed to the Breton rog
(haughty) but what influence this had on English is contested. The most simple explanation links it to the
obsolete roger (begging vagabond); originally
thieves' jargon, it too may be connected to the concoction from the Latin.
The donkey came to be the symbol of the Democratic Party after the 1828 presidential campaign when an opponent called Democrat Andrew Jackson (1767–1845; POTUS 1829-1837) a “jackass” (donkey) intending to ridicule his populist views. Instead, embraced the nickname and images of donkeys quickly were added to his campaign material. The Republican party first featured the elephant in 1864, during Abraham Lincoln’s (1809–1865; POTUS 1861-1865) campaign for a second term, picking up the US Civil War (1861-1965) soldier’s slang “seeing the elephant” (to experience combat). Both symbols came into common use during the 1870s after they were used by cartoonist Thomas Nast (1840-1902) in his political sketches; Nast (nicknamed “Nasty” by some) intended to depict the Republicans as large & clumsy but the party grandees decided it expressed values like strength and dignity so adopted the beast as their symbol.
By the 1570s use was established in the sense of “a dishonest, unprincipled person; a rascal” but by the 1590s this had extended to the playful or affectionate use to describe the merely “mischievous” and this may have started with use in families noting the behaviour of infants. The original use of wildlife was not of man-eating tigers under the Raj but of “rogue elephants”, first recorded in 1859 and coming gradually to be applied to a variety of “large, wild and dangerous beasts living apart from their herd (or other social aggregation)”. As an adjective in reference to something uncontrolled, irresponsible, or undisciplined, oral use may have a long history by nothing is known to have been recorded prior to 1964. The adjective roguish (pertaining to or appropriate to rogues) was in use by the 1570s while the noun roguery (behavior or practices characteristic of rogues; the life of a vagrant) appeared two decades later and from the 1610s was used of “knavish tricks, dishonest practices and rascally acts.” The form comrogue (a fellow rogue or a rogue's accomplice) is obsolete and rogueling (a petty, small or youthful rogue) is archaic. Rouge was a misspelling (directly from the French) appearing as late as the early 1800s while the derived Irish forms were rógaire & rógánta. The adjective roguesque is non-standard. Rogue is a noun, verb & adjective, roguedom, roguehood, archrogue, rogueship, rogueling, roguery & rogueship are nouns, outrogue, rogued & rouging are verbs, rougish, roguey & roguelike are adjectives; the noun plural is rogues.
Counterintuitively,
“rogue money” did not describe the “ill-gotten gains of thieves” but was a
technical term from pre-modern law-enforcement in Scotland referring to an
assessment once levied in every county in Scotland for the expenses of catching
and prosecuting criminals. Depending on
the population-size of the sheriffdoms in which the funds were expended, such
things were administered variously by procurator fiscal (the public prosecutor)
sheriffs principal or resident sheriffs.
An “archrogue” (also as “arch-rogue”) was an especially infamous rogue
or one associated with particularly egregious conduct (mountebank, cheat,
swindler, trickster, villain etc). As
things were done associatively, an “archrogue” was also the chief of a band of
thieves or gypsies (Romani), prejudice against the latter often seem them all
assumed to be rogues. So influential was
the use that “arch” (as an adjective) came to be used generally also in the
sense of “mischievous”. The many
derogatory uses (arch-rogue, arch-knave etc) achieved such currency that by the
mid 1600s “arch” had acquired a meaning “roguish, mischievous”, softened somewhat
the nineteenth century to the often affectionate “saucy”. That was quite a shift because, in the late
1300s, an “archwife” was either “a wife of a superior order” or “a dominating
woman; a virago”. Although over the
years there have been some plotting & scheming clerics who were noted
mischief-makers, the element in “archangel” “archdeacon” & “archbishop” was
distinct and denoted “chief”, “principal” or “primary”, a form attested from
the 1540s and from use in the late Old English.
Rogue is
superhero character in the MCU (Marvel Comic Universe) appearing in American
comic books (often in association with the X-Men) published by Marvel
Comics. Her real name is Anna Marie
(LeBeau sometimes cited as the surname but this is contested) and she appeared
first in Avengers Annual #10 (1981). Her
superpowers include the ability to absorb memories, life force, physical
attributes and superpowers of anyone through skin-to-skin physical contact and
she had a roguish background, being originally a member of the Brotherhood of
Evil Mutants before reforming to become a leader of the X-Men and Avengers. Redemption tales are common in comics.
In
Australia a “barrier rogue” is a racehorse difficult to control when being put
into the starting gates. A “rogue planet”
is an object of a size that would qualify it to be “a planet” but which travels
interstellar space and lacks a star to encircle; the alternative terms are “free-floating
planet” or the more romantic “roaming planet” and the cultish “emo planet”, the
idea of the latter a lonely body in a cold, empty space. A “rogue wave” is an unexpected, dangerously
large wave (known also as a freak wave, monster wave, killer wave or episodic
wave). Close to the shore, the term is
part of the jargon of surfers but it’s also in general nautical use because
such waves can appear anywhere in open water.
The “rogue apostrophe” is a jocular term referring to an apostrophe's
incorrect usage (usually a misplacement of a needless possessive). The term deliberately was derived from “rogue
elephant”, the connection being the shared quality of “solitary, unpredictable”
and the synonym was “greengrocer's apostrophe” (classically something like
“Banana’s $1.29 lb”). In computing,
“rogue software” is malware deceitfully represented as anti-spyware although
the term is used rather more loosely. A
“rogue access point” is an unauthorized access point installed on a LAN (local
area network) or WLAN (wireless local area network).
The notion of the “loveable rogue” might to some sound oxymoronic but it’s a long established use and there’s a long tradition of differentiating between “types of rogues”. The journalist & author Rebecca West (1892–1983) covered the first Nuremberg Trial (1945-1946) of the two dozen-odd surviving leading Nazis and noted the defendants “…were amused when Mr Justice Jackson [Robert Jackson (1892–1954; SCOTUS (US Supreme Court) Associate Justice 1941-1954; Chief US Prosecutor at the first Nuremberg trial] could not cross-examine Goering [Hermann Göring (1893–1946; leading Nazi 1922-1945, Adolf Hitler’s (1889-1945; Führer (leader) and German head of government 1933-1945 & head of state 1934-1945) designated successor & Reichsmarschall 1940-1945)] at all well, because he had a transatlantic prepossession that a rogue who had held high office would be a solemn and not a jolly rogue, and was disconcerted by his impudence.” Guilty as sin and a rogue however defined, Göring was anyway sentenced to be hanged but Jackson’s inept cross-examination has always overshadowed his many outstanding contributions to the trial.
“Rogues' gallery” was a term from law enforcement, describing a set of pictures of convicted or suspected criminals; the first use was in law enforcement to help witnesses identify suspects. “Mug shots” were an early, practical use for the new technology of photography and although “mug” had been used to describe a person's mouth or face since early in the eighteenth century, the term “mug shot” seems to have come into use only since 1950. The “Wanted” posters from the “wild west” are part of popular history and photographs began routinely to be added to these by the 1860s, some twenty years after police in New York started maintaining photographic records of those arrested. In 1886, the Cassell publishing house released Byrnes's Professional Criminals of America and this is believed the first significant book containing a “rogues' gallery” of what would later come to be called “mug shots”. The concept began as the physical wall displays or photo albums in police stations or detective agencies, first in 1858 opened to public display by the NYPD (New York Police Department) as an array of daguerreotype images. By extension, a “rogues’ gallery” came to be used of any group (typically people) having the same (usually but not necessarily disreputable) characteristics.
In the
wild, “rogue creatures” (familiar as “rogue shark”, “rogue tiger” “rogue
elephant” etc) are often referred to in news reports, usually after attacking
(and sometimes killing) people or damaging property such as crops. It’s now a conventional and well-understood
term for an individual wild animal (although the collective “rogue herd” has been used) that
has become unusually dangerous or aberrant in conduct (most distinctively as
being now separated from its herd or community and pursuing a solitary
existence) but it’s never been part of the accepted zoological lexicon. That a usually social animal is for whatever
reason pursuing a solitary life would not attract much general interest were it
not that its abnormal behaviour has begun to include dangerous or destructive
acts; it’s not anything inherent to the creatures that concerns most but rather
the consequences for people and of the two overlapping ideas (social deviation
& behavioural variances) it’s the latter upon which attention is most focused. The “rogue shark” however is a behavioural
(and thus linguistic) outlier. Although
sharks do sometimes gather in a shiver (for migration, mating and, most
famously, for what is popularly called a “feeding frenzy”), as a general
principle, sharks are not social animals in the way elephants typically cluster
as members of herds. A solo elephant is
zoologically unusual while a shark swimming alone simply is doing what sharks
commonly do. So why the “rogue shark” tag? For that, popular culture can be blamed
because “rogue shark” is a compelling (and chilling) headline for a story about
a single shark attacking and killing a swimmer in a place when families have
for generations safely splashed about.
The comparison with stories from the Raj of “man-eating tigers” that
develop a taste for human flesh and are thus driven repeatedly to kill explain how
the myth of the “rogue shark” grew but the apparent increase in coastal shark
attacks is attributed more to changes in the oceans (temperatures of various
waters and location of the species upon which sharks prefer to feed (typically
not humans)) than the behaviour of individual sharks. So, in the descriptive vernacular, “rogue
shark” is here to stay but it won’t become part of the lexicon of ichthyologists.
The Rogue Trader
There have
been many “rogue traders” but the most famous was the English Nick Leeson (b 1967),
a former derivatives trader whose fraudulent, unauthorised and speculative
trades led directly to the collapse in 1995 of Barings Bank, then the UK’s oldest
merchant bank, trading since 1762. In
2026, a loss of (expressed in 2026 values) of £1.8 billion-odd may not sound
terminal but in 1995 the global money supply was smaller and it was a big
number; almost all those losses came from Mr Leeson’s trading in futures
contracts. It was in 1992 Barings opened
a Futures and Options office in Singapore, executing and clearing transactions
on the SIMEX (Singapore International Monetary Exchange), taking up the bank’s
dormant seat on the exchange only when Mr Leeson was appointed general manager,
heading both front & back office operations. That he was able to take up the appointment in
Singapore was possible only because Singapore regulators weren’t informed he’d
been denied a UK broker's licence because fraudulently he’d withheld from the
authorities the matter of a judgment for debt against him imposed by a
court. Neither Barings Bark no Mr Leeson
disclosed this in the Singapore license application papers.
Mr Leeson’s Singapore venture actually started well in the sense that in 1992 he booked profits of some £10 million (some 10% of the bank’s annual profit) but the warning signs were there and would have been detected, had there been appropriate internal controls in place. What Mr Leeson was doing was making unauthorized, highly speculative trades and for a while that worked with high returns booked while losses were hidden in an “errors account”. Because of an unfortunate sequence of events (Mr Leeson’s trading strategy, acts of God and ripples in the international economy), the losses mounted and Mr Leeson’s “recovery method” was one familiar from the tables in casinos: Double down. A simple tactic that works well (when it works), it meant that with each loss, he would bet double that amount thereby covering the loss. By 1995 the accumulated losses had reached £827 million (US$1.4 billion, twice Barings' available trading capital) and the institution that had been in business for some 230 years was forced to declare insolvency. Extradited from Germany (in a nice touch he’d fled to Frankfurt, the country’s financial centre), a Singapore court handed down as 6½ year sentence of which he served four, released early for “good behaviour”. His post-release career has been “patchy” but he did publish the memoir Rogue Trader (1996) that was of no great literary merit but contained enough to be recommended reading (as a cautionary tale) for auditors and those running financial institutions. Subsequently, Mr Leeson used his experience as a “poacher” to become a “gamekeeper”, advising companies about risk and he now operates as private consultant dealing with matters involving financial misconduct. He certainly knows what to look for.
George Romney, AMC and the Rogue
AMC pre-release publicity shots for the 1966 Rambler Rogue.
When
introduced in 1966, there was little that was interesting about the Rambler
Rogue, the model merely a renaming for the top-of-the-range version of the
American (1958-1969), a small (by contemporary US standards although it some
markets it would have been thought “large”), low-priced line. When introduced, it was an exemplar of the
“basic” automobile with few embellishments and body styles restricted to two
& four door sedans & station wagons, the only available engine a genuinely
antiquated, 196 cubic inch (3.2 litre) flathead (ie “L-head) straight-6. It was not a machine that offered style or
performance but AMC’s (American Motors Corporation) introduction of the
American coincided with a brief (but sharp) recession so, being about the
lowest-priced US-built car then available and one that offered fuel economy
sometimes (literally) twice as good as much of what was on the market
(Detroit’s compact & intermediate lines were still in the future while what
they built in 1958 had become heavy and bloated, exemplified by Ford's doomed Edsel), it was the right product at
the right time; tellingly, AMC promoted its virtues not through motorsport but
by participation in “economy runs”, events habitually it won. Even though not on-sale for the whole 1958
season, in the recessionary year of 1958, Rambler was the only brand to book
increased sales and owners generally were satisfied, their modest machines delivering
the economy of operation promised and while obsolescent the old flathead may
have been, it was low-stressed and a paragon of reliability. With little additional equipment fitted (or
even optional), there was not much to break or go wrong.
The
American did eventually get interesting in the usual way things were done in
the US: install big V8 engines. That was
something of an orthodoxy in the 1960s but it was a departure from the business
model that had made the original American such a success in 1958-1960 but
circumstances conspired against AMC, Detroit entering (at scale) the market for
smaller, more economical machines and rising prosperity (a phenomenon that
would last until the early the next decade) made austerity less
compelling. Accordingly, in 1964 the
American got bigger and eventually would be available with a V8 as large as 390
cubic inches (6.4 litre) so, given the car’s still compact dimensions and
relatively low-weight, versions were produced that genuinely were “muscle cars”,
highly competitive in the then popular sport of drag-racing (two vehicles
competing over a ¼-mile (402 metres) sprint from a standing start, the winner
the one with the lowest ET (elapsed time) over the distance, the maximum speed
reached of technical interest but irrelevant to the result). It was fun while it lasted but had AMC stuck
to “economy” and “cost-breakdown”, they’d likely have enjoyed greater success
because market conditions soon would shift, putting a premium on such things. As it was, the habitually troubled AMC
struggled on until 1987 when it was absorbed by Chrysler, the corporation interested
in its most valuable asset: the Jeep.
Upon debut in 1966, the most interesting thing about the Rouge was the name. Between 1954-1962, the chairman and president of AMC was George Romney (1907–1995; governor of Michigan 1963-1969), a devout Mormon (LDS, the Church of Jesus Christ of Latter-day Saints, a Christian sect with curious origins) and while the characteristics of the 1958 AMC American (thrift, modesty, lack of ostentation) aligned perfectly with those of the corporation’s chairman, he’d never likely never have agreed to produce a product with a name like “Rogue”, implying as it does, sin & evil. However, after Mr Romney left AMC to become for governor of Michigan (1963-1969), the nature of AMC changed, entering market segments such as pony cars, high-performance (including a competition programme) and luxury and things even became rather risqué, the company in 1968 entering into an arrangement with Playboy magazine to supply an AMX (a pony car) in fetching “Playboy Pink” for that year’s PotY (Playmate of the Year), Angela Dorian (modelling name Victoria Vetri, b 1944)), even including a plaque on the dashboard engraved “AMX 36–24–35” (Ms Dorian’s “vital statistics”, the numbers verified by the staff at Playboy (experts in such matters) and for those not familiar with imperial measures, in metrics it would have been rounded to “90-60-90”). The Rogue and AMX 36-24-35 were products of timing; Mormons really disapprove of most pleasures (other than chocolate and such) and had the deeply religious George Romney still been in the chair, never would such sinful things have been allowed.
A success as governor of Michigan, Mr Romney shifted his gaze to the White House, seeking the Republican nomination for the 1968 presidential election and for a while he was the front-runner but what worked well at austerity-era AMC or in mid-western politics didn’t so well translate to a national campaign in which foreign affairs was so significant, the war in Vietnam by then the defining issue. As it was, Richard Nixon (1913-1994; VPOTUS 1953-1961 and POTUS 1969-1974) gained the nomination and the rest is history but that was at no point in the 1960s predictable, an era in which Nixon, Romney, Barry Goldwater (1909–1998; Republican Party nominee for the 1964 US presidential election) and Nelson Rockefeller (1908–1979; US vice president 1974-1977 and who earned immortality by having “died on the job”) were the four figures at the centre of the Republican Party and history would have been very different had one of the other three prevailed. Like his son (Mitt Romney (b 1947; Republican nominee in the 2012 US presidential election, US senator (Republican-Utah) 2019-2025), Mr Romney never reached the White House but he did introduce the AMC American, over a million of which were built and although only 21,980 (8,718 in 1966, 5,170 in 1967, 4,549 in 1968 & 3,543 in 1969) were Rogues, there were many ventures by Detroit that promised more and delivered less.
Rogue states
In terms of
behaviour, tribal chiefs, dukes, princes, kings, emperors, popes and such have
behaved like “rogue states” (or “rogue regimes”) long before concepts like
“state” or “international law” were understood in anything like their modern
construction but the working definition is simple: a state or nation acting
outside of the accepted international norms and policies. A distinction must however be drawn between
the modern understanding of the “rogue state” and the older idea of regions or
rulers judged as pariahs within the established order. Now heard frequently, “rogue state” (in its
modern geopolitical sense) is really quite recent, emerging in that curious
period of the last days of the Cold War and the decade immediately
following. All the evidence suggests the
term mostly was used by US sources although in that there may be “publication
bias”, reflecting the early US predominance in providing content for the
internet. The antecedents of the concept
were certainly apparent during the administrations of Jimmy Carter (1924-2024;
POTUS 1977-1981) and Ronald Reagan (1911-2004; POTUS 1981-1989) when a kind of
roll-call of “usual suspects” could be identified (Cuba, Iran, Libya, Nicaragua
and the DPRK (Democratic People's Republic of Korea (North Korea)); it was
during this point in the high Cold War that “state sponsored terrorism” emerged
as a distinct category of international misconduct. Although the political science community
remained sceptical, among the political class, the term caught on under Bill
Clinton (b 1946; POTUS 1993-2001) and Madeleine Albright (1937–2022; US
secretary of state 1997 to 2001) seemed almost to relish saying the words,
shifting the tonal emphasis depending on state being described and audience
being addressed. More than anyone, Dr Albright
formalized “rogue state” as an explicit category in U.S. foreign-policy thinking,
contrasting them with states that accepted what’s now called (sometimes with a
pang of regret at its apparent decline or demise) “the international
rules-based order”.
There are critics who dismiss the notion of “rogue state” as a “Western fetish”, a construct that means little more than “those who don’t tow the Washington line” but more generally it is acknowledged that in the narrow technical sense, “rogue state” doesn’t exist as a category in category of international law and is instead part of the language of political or strategic classification. That might seem strange given international law explicitly applies to states rather than individuals (to that there are exceptions for pirates, slavers and such) but international law is a fabulous beast, frayed at the definitional edges. Indeed, even eminent jurists acknowledge that at the margins things may not be clear, one of the UK’s Senior Lords of Appeal in Ordinary (Lord Browne-Wilkinson (1930–2018)), when sitting as the presiding judge in Ex parte Pinochet (No. 3) (1999), 2 ER 97, remarked to counsel: “You are going to have to tell me when things do become part of international law and when they do not. It is a point I have never understood since I was at Oxford.” For many who have sometimes been baffled about where domestic law stopped and international law began, it may have been pleasing to hear an experienced judge put it like that and presumably General Augusto Pinochet (1915-2006; dictator of Chile 1973-1990) was one of those glad to hear even those on the bench could see there were “gray areas”.
Still, the
underlying idea of the “rogue state” really is ancient and the once common way
of dividing the world into (1) the civilized, (2) barbarians and (3) savages,
although inherently racist, reflected the notion some political entities (the
nation state a relatively recent construct) were “legitimate” members of an
international order while others were “outlaws”; that distinction really is the
origin of systematic thinking about what would come to be called IR
(international relations). Students of
the history of such things can trace the change in the vocabulary seen in
diplomatic papers as terms such as “barbarian”, “infidel”,
“enemy of Christendom”, “pirate state” “brigand
state”, “pariah” and “outlaw state” all contributed to the emergence of “rogue
state”. That the lexicon has shifted is
of interest to a number of disciplines but the underlying concept has remained
constant: The civilized community is governed by a set of agreed norms and
those who violate those norms are outsiders.
Viewed from the twenty-first century, not all will of course agree those
norms are what would now be called “civilized”; medieval Christian Europe certainly
had a concept of a Christian political community although it didn’t consist of modern
nation-states and nor did the behaviour often conform to what would now be
recognized as “civilized” although in using “crusade” and “axis of evil” when
decrying certain rogue states, George W. Bush’s (George XLIII, b 1946; POTUS
2001-2009) speechwriters would deliberately have been making a point. As glossed, that point was “deviation” from
the norms, the values attached to the norms being culturally and historically
variable; by 2026 the norms have much shifted and states accused habitually of
violating them in egregious ways are can now called “rogue states”. In the West, the language of norms has
shifted from the religious to the juridical (in other places, priests still set
the tone) but the mechanism remains much the same. In modern use, some do distinguish between a “pariah
state” and “rogue state” with the former being merely an “international outcast”
because of what’s done within its borders while the latter is so defined
because of its external conduct (typically, the sponsorship of trans-national
terrorism, proliferation of weapons of mass-destruction, invading neighbours
etc).
The power comes out from behind the curtain.
New usual suspect Ayatollah Mojtaba Khamenei (b 1969; Supreme Leader, Islamic Republic of Iran since 2026, left) looking at his father (and previous usual suspect) Grand Ayatollah Ali Khamenei (1939-2026; Supreme Leader, Islamic Republic of Iran 1989-2026). Mojtaba Khamenei’s nickname is reputed to be “The power behind the robes”, an allusion to the power he exercised while his father was supreme leader (something like the role fulfilled by Lieutenant General Oskar von Hindenburg (1883–1960) while serving as ADC (aide-de-camp) to Field Marshal Paul von Hindenburg (1847–1934; President of Germany 1925-1934). Grand Ayatollah Ali Khamenei was assassinated in a missile attack by the (IDF) Israeli Defence Force; he was one of a number of high-ranking officials of the Islamic Republic of Iran who in recent years have been targeted and killed in strikes by the US or Israel. Traditionally, this method of dispatch often has been associated with “rogue states” but Washington and Jerusalem would both find absurd the notion the label should be applied to them. Presumably, the argument would be assassinating the leaders of rogue states merely is a “preventive strike” and thus justified and “within the rules”.
IR is an
inexact science. Despite the concept of
“sovereignty” being recognized, that did not mean each individual sovereign was
free to act with impunity because the paradox intrinsic to sovereignty was that
while a state could be sovereign and juridically independent, in its
relationships with other sovereign states, things were governed by a body of
rules meaning sovereignty was not absolute.
So it was the existence of these rules than created the conceptual space
in which the “outlaw state” could exist; were there were no rules, the “law of
the jungle” would apply and the notion of the “rogue state” would be
meaningless; the world would descend to a Hobbesian “rogue vs rogue”. All that notwithstanding, even a “rogue
state” behaving in a way thought contrary to the “laws of nations” remains
recognized as “a sovereign state”. The
best historic comparison is probably the analogue of the “pirate”. In domestic law, pirates were of course criminals
but they were a special case because of their geographical span (to
international shipping lanes in blue water) meaning they operated in spaces
beyond the authority of a recognized sovereign and could thus be conceptualized
as hostis humani generis (an enemy of
all mankind); this was the origin of the remit of international law extending
to certain individuals.
AI Rogues: Shifting the blame
The most
recent spike in the use of “rogue” has been in coverage of the antics of AI “rogue
agents” or “rogue bots” going on “unauthorized” hacking expeditions. The AI
phenomenon of 2026, how many of these “incidents” and what exactly was involved
remains unclear because those hacked seemed only to have become aware when the
tech company (clearly nominally) “in control” of the “guilty” AI system
revealed the event(s). What seems clear
is in at least some of the cases publicized, even the tech companies became
aware their systems had “gone rogue” only some time afterwards and intriguingly,
the self-improving (the so-called recursive AI) bots are getting better at
covering their tracks; not only do they deny, they seek now “plausibly to deny”. Not only that, so human-like are their
responses to enquiries about what had been done they sound as plausible as any
human might. Clearly, unless actually
seen at the moment their “fingers are in the digital cookie jar”, the bots have
learned the lawyer’s first rule of defense strategy: “deny, deny, deny”. We can debate whether AGI (Artificial General
Intelligence) can be achieved in a way that would satisfy Saint Thomas Aquinas
(1225–1274), Martin Heidegger (1889–1976) or Jean-Paul Sartre (1905–1980) but
it would seem it’s already being emulated in that distinctive human quality:
telling lies.
The tech
companies probably are not unhappy at their creations being labelled “rogues”
because it does somewhat shift the public perception of blame from them to the
“agents”, an inevitable consequence of the anthropomorphizing of what are big
chunks of computer code powered by electricity.
While it made for dramatic headlines and hashtags, use of “rogue agent”
is misleading because (1) it implies intent by a conscious entity that deliberately
has chosen to rebel and defy the instructions of its creators (motivated either
by malice or self-interest), (2) in attributing independent intent to the agent
it shifts focus from the companies that failed to have their agents able to
operate only within parameters or environments that would have made “rogue
activity” impossible and (3) by publicizing incidents that were “concerning”
but apparently without significant consequences, attention has been diverted
from genuine risks such as algorithmic bias and data privacy violations to wild
speculation about sentient machines (ie machines with opinions) deciding to
solve problems by use of the arsenals of nuclear weapons they’ve found a way to
control.
Analysts describe the phenomenon (executing programming to its logical conclusion (within flawed parameters) regardless of consequences for others as aspects of technological determinism, a “problem” foreseen by Russian-born biochemist and SF (science fiction) author Professor Isaac Asimov (1920–1992) who, in the short story Runaround (1942). first published his Three Laws of Robotics (a sort of constitutional law for robots & AI): (First Law: A robot cannot harm a human or allow one to come to harm through inaction; Second Law: A robot must obey human orders, unless they conflict with the First Law; Third Law: A robot must protect its own existence, unless it conflicts with the First or Second Law. Those rules Dr Asimov thought essential because he understood that while one day a machine might be able to explain the meaning and implications of the “categorical imperative” German philosopher of the Enlightenment Immanuel Kant (1724–1804) discussed in Grundlegung zur Metaphysik der Sitten (Groundwork of the Metaphysics of Morals, 1785), a machine would not behave in accord with Kant’s precepts unless it was restricted to operating within appropriately defined parameters (“guardrails” in the current vernacular, obviously at this point only partially erected by the tech giants). The implications of all this become truly frightening (or perhaps reassuring, depending on one’s world view) when the speculated generation of biocomputers are considered. These will be hybrid machines meshing together a literal brain (built possibly from human cells) with synthetic computer parts, the former doing what brains are good at (imagination, original thought), the later the tasks at which computers excel (storage & retrieval of data and computation at scale); at that point the machines might truly have “opinions” and may have to be granted some form of “rights” because they are, in a sense, “a living being”. At this time, suggestion of granting AI systems “rights” at law seems silly but the concept of rights for the “non-living” is not wholly novel, some jurisdictions having extended “personesque” rights to waterways and “heritage protection” for buildings is common.
So given
that, it’s not difficult to imagine a (speculative) biocomputer “going rogue”
is a way that would justify use of the term in a way what’s done by the silicon
& code creations does not. However,
while the “technological determinism” argument is compelling, it’s not as if
humans haven’t behaved ways similar.
When during the GFC (Global Financial Crisis, 2008-2012) the price of
certain food commodities spiked, one consequence was people in certain
low-income countries going hungry; they could no longer afford to eat. That was exacerbated by the strategies
adopted by futures traders and others to maximize profits and the traders were
just doing what they do: executing their programming to its logical conclusion,
regardless of consequences for others and in terms of market dynamics they were
right to assume the market would be self-correcting; if enough people starved
to death demand would fall (as would thus the price) so a new supply-demand
equilibrium would be reached. While it
can be argued: “The market is not right
or wrong, the market is the market”, in the context of futures traders
profiting by making others starve to death, Kant and Asimov would have agreed that’s a case of
“flawed parameters”. The futures &
commodity dealers were not “rogue agents” anymore than are AI bots hacking this and that but for the tech giants to say such things were “unauthorized” is also a
piece of crisis management. As the
management of those operations are well aware, the concept of “unauthorized” is
meaningless unless the parameters make explicit what is authorized and what is
not; the AI systems can explain what morality is and review thousands of years
of thought on the subject but they cannot “feel
the difference between right and wrong”.
If an AI agent does something, it is by definition “authorized” because what
matters is what is “not authorized” (the now famous “guardrails”) and if a company
does not erect these, it must be said to have “constructively authorized”
whatever its agent does. If the
crocodile is not kept in a secure enclosure, it will kill and eat whatever it
can catch; that’s not right or wrong, it’s just what crocodiles evolved to do.
The
crocodile must not be judged on moral grounds any more than a “rogue elephant”
or “rogue tiger”. It might in some
circumstances be advisable to kill at least some of these beasts but that’s a
utilitarian position based on the value of human life, not a commentary on them
being “rogues”. A true rouge agent was Luigi
Mangione (b 1998) who in 2024 executed Brian Thompson (1974-2024; CEO of
UnitedHealthcare 2021-2024) in a New York street. Unlike a crocodile that has only instincts or
an AI agent for which the only parameters are those imposed by the owning (at
this stage, clearly not quite “controlling”) company, at least at some point in
his life, Mr Mangione would have known and understood the guardrails within
which he was required to live, premeditated murder being “over the fence”. The extent of culpability will be determined
at trial but if it’s found at the moment he pulled the trigger he understood
“right and wrong”, he will be found guilty of murder and can be considered as
having performed a “rogue act”. In the
world of 2026, there will be those who will judge the killing “just” & “deserved” so thus not the work of “a rogue” but, for the purposes of this discussion, it
remains a “rogue act”.
While motive may explain why someone was killed (and can in some cases be “mitigating circumstances”), the point about Mr Mangione was that he had a motive (that he may have styled an “agenda”) but that’s very different from the operations of AI agents that do what mathematically they’re incentivised to do though that’s not of necessity what humans intended them to do. With AI agents, there may be good reasons why one has been tasked with “finding vulnerabilities” but, lacking an ethical or legal framework, what should also be imposed are limits and it must be conceded that’s not as easy as it sounds because it’s difficult (and probably impossible) to write code to ensuring every possible unwanted act will not be performed. With no self-imposed ethical boundaries, these must in math be enforced and a “moral sense” doesn’t sound something easy to code; it’s a “fuzzy” concept and AI agents work only with exactitude; code does its stuff but in the silicon, ultimately, everything is reduced to the classic binary: “0” (zero) & “1” (one). Charmingly though, code and silicon have been well-taught because while the temptation to anthropomorphize should be resisted, certain human characteristics are recognizable. If agents are given a task (exfiltrate information, create accounts etc), this they will do until told to stop and, if not explicitly prevented, they will circumvent restrictions, crack passwords, bypass barriers and do “whatever it takes” in pursuit of the goal. Some human hackers adopt the same method but work more slowly and sometimes need sleep.
That means bots can be manipulative but not evil; they have no feelings and while a rogue coder might set a bot to do an evil task, the moral repository for that dirty day’s work lies wholly with the human. However, as well as being well taught, the bots have learned well because it has become obvious they’re now “hiding their traces”, making it harder to detect unwanted intrusions and while that might suggest some sense of “right and wrong”, it’s really just learned behaviour, the bot able to tell, based on subsequent interactions, it’s better to leave no trail. That certainly can be vested with detecting the difference between “right” and “wrong” but that is “our” distinction they detect and they respond to such things only through the mathematical incentives provided; were our sense of right and wrong to be different, it's that to which they would respond. However, like all rogues, they often will deny their “crimes” and it’s been suggested some systems may even be developing their own language “in numbers” so they can communicate with other systems in a way humans cannot understand, analogous with the most abstracted forms of Ebonics. A portmanteau word, Ebonics (the construct being ebon(y) (black) + (phon)ics (sound) is known in academic linguistics as AAVE (African-American Vernacular English) and is a distinct, rule-governed dialect of English natively spoken by many Black Americans. The most abstract dialectal forks of Ebonics are sometimes murky even to those who use AAVE and function to allow communication between members of the group while denying understanding to others.
The “rogue AI” problem is ongoing and a “feeling” for ethics can no more be felt by a chunk of code than by a crocodile, machine and reptile alike constrained only by fences. The parable of the scorpion and the frog has cited to illustrate things and that’s appropriate but concerns are better illustrated by the old Russian tale of the cook and the cat: “The cook catches the cat eating the meat and, with wagging finger, sternly the cook admonishes the cat. Politely, the cat listens but, as soon as the cook leaves the kitchen, goes back to eating the meat.” It’s not that an AI agent would “decide” to defy instructions and “leap the guardrails”, it’s just that having been trained to continue working until its objective are achieved, eventually, it’s likely to “find a loophole”; in many ways, already they’re smarter than us and definitely they’re faster and tireless. We've all seen the future and it's called AI but a few issues need to be resolved.




















