When Criticism Becomes a Licence to Hate

Charlie Kirk was shot dead while speaking at Utah Valley University in September 2025; after the 7 October attacks and Israel’s war in Gaza, Jewish students at American universities reported harassment and exclusion for perceived ties to Israel.¹⁶ The shared question is not whether public figures or governments may be condemned, but what happens when accusation is stretched until people who did not commit the alleged wrong become legitimate targets.

A bullet made into a verdict

The facts of Charlie Kirk’s killing are stark. At about 12.20pm on 10 September 2025, he was speaking at a Turning Point USA event organised by students at Utah Valley University. About 3,000 people attended; six university police officers were assigned to the event. Kirk was taken to hospital and later pronounced dead. It was a political speaker, addressing a political audience, killed in a public place where disagreement was part of the event’s purpose.

That fact should set the limits of what can be said next. Kirk was not an uncontroversial figure suddenly transformed into a martyr by his death. He was a prominent organiser whose arguments on abortion, race, religion and national politics provoked real opposition. His critics are entitled to examine those arguments, describe their consequences and say they were wrong. His supporters are entitled to mourn him. Neither right requires the other to pretend that the dead man was beyond criticism.

But public reactions to a killing often demand a faster verdict than the evidence allows. If a person is first presented as dangerous, hateful or beyond redemption, the murder can be recast as a kind of moral punctuation: a terrible event, perhaps, but one that happened to someone who had already forfeited ordinary sympathy. In that story, the victim’s death becomes proof of the accusation. The accused person is no longer someone whose views should be argued against; he is the threat the argument says he was.

This is the political habit at issue here: accusation inflation. It does not mean that accusations are usually false or that people should soften every charge. It means taking a claim beyond what the evidence supports—turning an offensive remark into a call for murder, an allegation into a verdict, or a government’s conduct into guilt attaching to an entire people. Inflated claims make the emotional work of punishment easier. They can also destroy the credibility of real criticism by teaching audiences that the facts are optional if the target is unpopular enough.

Stephen King deleted an X post and issued multiple public apologies after falsely claiming that conservative activist Charlie Kirk, who was assassinated on September 10, 2025, while speaking at Utah Valley University, had “advocated stoning gays to death.” The remark, posted the day after the shooting in reply to a tribute calling Kirk a patriot rather than a polarising figure, referred to a 2024 podcast clip in which Kirk quoted Leviticus to criticise children’s YouTuber Ms Rachel for citing “love thy neighbour” in support of Pride while ignoring a nearby verse that prescribed death for certain same-sex acts; Kirk was highlighting what he called cherry-picking, not calling for violence, and had previously said gay people were welcome in the conservative movement. After Senator Ted Cruz and others labelled the post a lie, King wrote that he had not fact-checked what he read on social media, conceded “Charlie Kirk never advocated stoning gays to death,” and restated that Kirk had merely demonstrated selective use of Scripture.

The distinction is not abstruse. Refusing to mourn a politician is not the same as celebrating his killing. Saying that a public figure’s policies harmed people is not the same as saying he deserved to die. Criticising a state is not the same as blaming every citizen, co-religionist or student associated with it. The trouble begins when people treat these distinctions as evasions, and when they portray the demand to make them as evidence of sympathy for the accused.

The false sentence and the true one

Two claims about Kirk show why precision matters even when a target’s actual views are open to severe criticism. One widely shared allegation said he had advocated stoning gay people to death. Snopes traced the claim to a podcast exchange in which Kirk quoted biblical passages while challenging a children’s entertainer’s use of scripture. The fact-check found that he had not directly called for gay people to be stoned. It also recorded the context that complicates any easy defence: Kirk called the section of scripture involving punishment “God’s perfect law” on sexual matters. The false allegation is not needed to make the real statement disturbing.²

That distinction is easy to lose online. “He cited a biblical passage in an argument about scripture” takes time to explain. “He wanted gay people killed” compresses the argument into a clean, shocking sentence. The compressed version is more shareable, more morally satisfying and less accurate. A reader who accepts it has not merely misunderstood the remark; the reader has been handed a different claim, one that invites a different judgement about the speaker.

A second example shows another route to the same result. Kirk was asked in a 2024 debate what he would want if a ten-year-old daughter became pregnant after rape. He said he would want the baby delivered, while arguing that the unborn child’s rights should not depend on how it was conceived. The hypothetical is real, and the view is severe. Viral versions sometimes changed the child’s age to nine. That adjustment does not make the underlying position more defensible; it makes a genuine, already shocking statement less exact.³

There is no journalistic virtue in correcting the age and then treating the correction as a defence of the policy. Accuracy is not absolution. Nor does a fact-check’s finding that one allegation is false certify the speaker as fair-minded or humane. It establishes only that the allegation, as phrased, does not fit the record. That limited result matters precisely because it does not settle every other question about the speaker.

In the first case, an inaccurate paraphrase makes Kirk more monstrous than the evidence allows. In the second, a real position needs no embellishment. Together they expose a recurrent political temptation: to believe that if the underlying grievance is strong enough, a small distortion is harmless. It is not. Once the distortion is admitted, an opponent can use it to discredit the whole grievance. More gravely, the audience learns that a target’s supposed guilt matters more than whether the words attributed to him are his.

The International Court is not a verdict machine

The legal dispute in The Hague demands the same discipline, though the stakes are vastly greater. South Africa instituted proceedings against Israel at the International Court of Justice on 29 December 2023, alleging violations of the Genocide Convention in Gaza. The case has produced provisional measures and a continuing legal process. It has not produced a final ruling that genocide occurred. The Court’s current case file records further procedural steps in 2026, including an order setting deadlines for written pleadings that run to 2027 and 2029.⁴

Provisional measures are not a verdict in abbreviated form. They are interim directions intended to protect rights while a case is pending. In its order of 26 January 2024, the Court concluded that at least some rights claimed by Palestinians in Gaza under the Convention were plausible and that the conditions for provisional measures had been met. It directed Israel to take measures to prevent acts covered by the Convention, prevent and punish direct and public incitement to genocide, enable urgently needed humanitarian assistance and preserve evidence. Those are consequential orders. They are not a finding that Israel committed genocide.

The difference between “plausible rights” and “genocide proved” is not a lawyer’s trick. The Court was considering whether to protect rights before it had heard the case in full. It did not make a final determination of the facts, intent or legal responsibility. International law’s threshold language matters because the Convention defines a grave crime requiring proof of particular acts and a specific intent to destroy, in whole or in part, a protected group. Treating an interim threshold as a completed finding makes a court say something it has not said.

The first order also did not grant South Africa’s request for a general suspension of military operations. That omission is relevant, but it is not a ruling that the accusation was baseless. Later orders addressed the deteriorating humanitarian situation. On 28 March the Court added measures concerning urgently needed aid; on 24 May it reaffirmed earlier directions and ordered Israel to halt its military offensive in Rafah, and any other action there, insofar as it could inflict conditions of life capable of bringing about the physical destruction of Palestinians in Gaza, in whole or in part. A concise account must preserve both what the Court ordered and what it did not.⁵

Those details resist slogans from either camp. “The Court found genocide” is false. “The Court threw the case out” is false. “No ceasefire order means the case is a hoax” is also false; the Court issued binding provisional measures and kept the merits question open. “The measures prove genocide” is equally false. Political urgency may explain why people want a final answer now. It cannot supply one on the Court’s behalf.

The record is still open

The legal process has become a global political object. For some campaigners, the word “genocide” now functions less as a charge awaiting adjudication than as a settled description of Israel and, by implication, anyone who refuses to repeat it. For some defenders of Israel, the opposite manoeuvre is tempting: because the Court has not decided the case, the claim can be dismissed as a political performance. Neither conclusion follows from the procedural record.

At the time of writing, the ICJ has set the next phases of written argument on a long timetable. Its order of 21 May 2026 required South Africa to submit a Reply by 22 November 2027 and Israel to file a Rejoinder by 22 May 2029. The Court’s case file also records interventions by other states and further developments in 2026. A case that has not reached a final decision should be described as unfinished, even if its provisional measures carry legal force and its factual subject is urgent.

That careful wording has a human cost. Gaza is not a hypothetical, and the deaths, displacement and destruction that gave the case its urgency are not erased by saying the Court has not yet ruled on the charge of genocide. The opposite is true as well: the seriousness of those events does not authorise a writer, politician or protester to report a judicial finding that does not exist. The charge is serious enough to require exact language, not less exact language.

International courts have limited power. They depend on states to comply with orders, and on governments to accept the authority of a judgment when it goes against them. That weakness gives every political faction an incentive to treat the Court as powerful when its orders support their cause and irrelevant when they do not. But if a court is to mean anything, parties and observers must distinguish an interim order from a judgment on the merits, and legal duty from political preference.

None of this is an argument for waiting silently until 2029. States, journalists and citizens can demand compliance with the Court’s existing orders, scrutinise military conduct and call for an end to civilian suffering. They can argue that the evidence already supports a stronger conclusion than the Court has reached. What they cannot honestly say is that the judges have already reached it. The central test is simple: can the accusation be stated in a way that survives contact with the actual decision?

The Russian comparison is useful because it is imperfect

The comparison between Jewish students after 7 October and Russian students after Russia’s full-scale invasion of Ukraine helps expose the difference between criticism of a government and suspicion of a diaspora. It does not work as a controlled experiment. The wars, university cultures, histories and political vocabularies differ. The circumstances of a Russian student in Berlin in 2022 were not identical to those of a Jewish student at Columbia or Harvard in 2024. Any argument that says they were identical has already made the comparison do too much.

Nor were Russian students left untouched by the response to the invasion. In March 2022, Germany’s Federal Association of International Students reported a rise in attacks, insults and threats against people presumed to be Russian, including in student residences and in public. It criticised measures that punished individual students for the actions of the Russian state, while supporting suspension of institutional cooperation with Russian universities whose leaders backed the war. The principle was clear: isolate institutions and officials where justified; do not assign every Russian student responsibility for the Kremlin.⁸

That response was not flawless, and isolated cases of discrimination matter. A Russian student who was insulted or denied an opportunity because of nationality did not experience an abstract policy debate; that person experienced collective blame in daily life. The comparison should not be used to dismiss those cases. Instead, it asks whether western institutions and campus movements generally treated Russian identity itself as a political offence, or whether they more often directed their opposition at the Russian state and its war.

After October 7, the boundary between state and person proved less stable for some Jewish students. A Jewish student who supported Israel’s existence, or simply would not disavow it on demand, could be treated as a representative of an army or cabinet. That is not the same as condemning an Israeli government. It is a change of target—from the state’s conduct to a student’s identity, presumed allegiance or right to take part in campus life.

The comparison therefore supports a limited, defensible claim: a society can oppose an aggressive state without treating every person of the state’s nationality or ethnic background as an accomplice. Evidence of discrimination against Russians shows that the principle was not applied perfectly. The accounts from Jewish students show that, in some American campus settings, it was applied unevenly. That is a serious criticism, but not proof that every protester was antisemitic or that criticism of Israel was illegitimate.

When students become stand-ins for a state

Columbia University’s own task force provides a documented view of what Jewish and Israeli students said they experienced after the attacks of 7 October. It reported that nearly 500 students gave testimonials in listening sessions. The accounts included harassment, verbal abuse, ostracism and, in some cases, physical violence. Students described being told to “go back to Poland”, feeling unsafe wearing visible signs of Jewish identity, or being excluded from student groups because of Zionist beliefs.

Those are testimonies collected by an internal university task force, not findings from a court after adversarial testing. The report said its examples were illustrative rather than exhaustive, and that the task force was not adjudicating individual complaints. That distinction should be preserved. It does not make the testimony worthless; it tells readers how to use it. The report establishes that hundreds of students brought accounts to the university and that its task force saw repeated patterns. It does not establish that every allegation was verified or that every student at Columbia shared the same experience.⁶

Harvard’s task force, which reported in April 2025, recorded similar concerns about Jewish and Israeli students being asked to distance themselves from Israel or explain their views before they could be treated as individuals. Its report also acknowledged limits in its method: the task force did not hear from every perspective, and some accounts were not independently corroborated. That candour matters. A serious institution can report a troubling pattern without claiming that its report is a criminal verdict.

The Harvard account is especially revealing because it describes turning an ordinary personal history into a political test. A Jewish student’s planned remarks about a grandfather who helped refugees reach what is now Israel were challenged by student organisers as politically unacceptable, according to the report. The task force said some Jewish affiliates were asked to prove they were “one of the good ones” by denouncing Israel. The phrase describes a climate, not a universal description of Harvard; it captures the pressure to make a private identity answer for a foreign government.

Harvard also commissioned a separate task force on anti-Muslim, anti-Arab and anti-Palestinian bias. That matters because prejudice on campus is not a competition in which one group’s evidence cancels another’s. Muslim, Arab, Palestinian and pro-Palestinian students also reported hostility and fear during the same period. A university that listens only to Jewish students fails; so does one that treats Jewish students’ claims as political camouflage. The standard must be the same: hear the account, investigate where possible, distinguish criticism from harassment, and do not make a student answer for a government.⁷

“Zionist” cannot be a substitute for “Israeli” or “Jew”

Political language does not merely describe a target. It can redraw the boundary around who counts as one. “Israeli” is a nationality; “Jewish” is a religious, ethnic and cultural identity; “Zionist” describes a range of political commitments connected to Jewish self-determination and the existence of Israel. The terms overlap in many lives, but they are not interchangeable. Treating them as synonyms lets a movement move from condemning a government to policing the identities of people who may never have lived in Israel, voted in its elections or served in its military.

There is no single Jewish position on Zionism, Israeli policy or the war. Some Jews reject Zionism; some identify with it while opposing the current government; some support Israel’s existence and sharply criticise its conduct; some are Israeli citizens with no simple view of the question. A slogan that treats every Jew as a Zionist is false. A slogan that treats every Zionist as an agent of the Israeli state is also false. People can be wrong about politics without ceasing to be individual members of a campus community.

At the same time, the boundary cannot be protected by declaring that any criticism of Zionism or Israel is antisemitic. That would inflate a different accusation and weaken the ability to identify real antisemitism. Students and scholars may argue that Israel’s policies are unjust, that its military campaign violates international law, that it should face sanctions or that Zionism as a national movement is flawed. Those arguments can be fierce and still be arguments about a state, a political ideology or a government.

The test changes when the demand is no longer “account for this policy” but “prove that you are not contaminated by the people, state or history with which I associate you”. Excluding a student from a club because she is Jewish, or demanding that she denounce Israel before she can speak, is not made legitimate by attaching the label “anti-Zionist”. Conversely, a student who opposes Israel’s existence cannot be assumed to hate Jews without examining what the student actually says and does.

This is where the Russian comparison becomes morally useful. A university can suspend a partnership with a Russian institution because its leadership supports an unlawful war, while refusing to make a Russian student personally answer for that institution. It can criticise Israel’s campaign and still refuse to make Jewish students answer for the Israeli state. The first principle is not softness toward either government. It is a refusal to shift blame from actors with power to people who share a name, religion or presumed political allegiance.

The street has evidence; it does not supply a complete causal theory

Britain’s antisemitism figures establish the scale of the problem, but they do not establish a single cause. The Community Security Trust recorded 3,700 antisemitic incidents in 2025, the second-highest annual total in its records and a 4 per cent increase on 2024. The total was below the 4,298 recorded in 2023, the year of Hamas’s attack and the start of the Gaza war. The figures are reports received by the Trust, not a census of every incident and not a count of prosecutions or proven crimes.

The detail makes the pattern harder to dismiss as an impression. Online reports reached 1,541, or 42 per cent of the annual total. October was the worst month, with 463 incidents. On 2 October, an attacker targeted people near Heaton Park Synagogue in Manchester on Yom Kippur; Melvin Cravitz and Adrian Daulby were killed and three others seriously injured. CST described the attack as the first antisemitic terrorist attack in Britain since its records began in 1984 to result in a loss of life. In December, 15 people were killed at a Hanukkah event at Bondi Beach in Sydney, an attack Australian police described as terrorism.⁹

The Trust’s analysis also complicates a simple equation between criticism of Israel and antisemitism. It reported that 1,977 incidents—53 per cent of its 2025 total—referenced or were linked to Israel, Palestine, Hamas’s attack or the subsequent war, and that each of those incidents also showed anti-Jewish language, motivation or targeting. A reference to Israel alone was not enough for CST to classify an incident as antisemitic. That methodological point should be stated plainly, because it rules out the misleading claim that the report simply counts criticism of Israeli policy as hate.

Tell MAMA, a service that monitors anti-Muslim incidents, recorded 1,909 reports in London in 2025, its highest annual total, and 1,031 in the first half of 2026. Those figures are not directly comparable with CST’s national totals: one measures London reports, the other incidents reported across the UK, and the organisations use distinct processes. They do show that anti-Muslim hostility is also rising sharply in a large British city. The answer is not to call one group’s experience evidence against the other’s. It is to refuse the politics that turns a distant war into permission to target local neighbours.

Neither dataset can establish that a particular court filing, protest, speech or news story caused a particular assault. The attack at Heaton Park cannot be explained by the ICJ case; Bondi cannot be reduced to a slogan from a campus encampment. But the data does establish that incidents rose in a period when war and Israel-related language dominated public debate, and CST found that many reported incidents involved both regional references and anti-Jewish content. That is evidence of a pattern worth investigating, not proof of a single causal chain. Anyone who claims more should show the evidence, not demand that readers infer it from timing alone.

South Africa’s record is relevant—and not a defence for Israel

South Africa’s decision to bring the case at The Hague gives critics a legitimate question to ask about its own treatment of minorities. Human Rights Watch documented attacks on non-national residents in South Africa, including killings, serious injuries, forced displacement and repeated failures to investigate or hold perpetrators to account. Its 2020 report described migrants blamed for unemployment, crime and economic hardship, and documented instances of officials and police failing to protect victims or contributing to discrimination. South Africa brought the case only weeks after October 7th 2023: the case was recorded on December 29th 2023!

That history is not a legal answer to the Gaza case. A government does not lose the right to invoke a treaty because it has failed to protect people at home. The wrong done to migrants in Johannesburg does not make Palestinian civilians less deserving of protection; nor does South Africa’s hypocrisy, where it exists, decide what Israeli officials or forces did. Using a state’s domestic failures as a substitute for examining the evidence in the case changes the subject while pretending to answer it.¹⁰

Yet the question of consistency remains politically fair. A state that appeals to universal duties should be asked whether it applies those duties to people who are unpopular or foreign at home. A movement that demands protection for one group should be asked whether it extends the same protection to groups it dislikes. That scrutiny is not whataboutism when it is kept in its lane: it tests the accuser’s credibility and moral consistency, not the truth of the underlying allegation.

It cannot settle the case by insinuation. “South Africa has xenophobia, therefore Israel did not commit genocide” is a non sequitur. “South Africa invokes universal rules, therefore its own record is beyond criticism” is just as weak. The accuser’s record belongs in the public account, but it cannot replace evidence about the accused. The same is true at individual scale: a person’s hypocrisy may be worth exposing, but it does not prove that every charge against that person is false.

The distinction matters because accusation inflation often presents itself as a demand for consistency while quietly substituting one question for another. Is the accusation true? Is the accuser consistent? Are victims protected at home? These are related political questions, but they are not interchangeable. A rigorous argument can ask all three and resist the impulse to use the easiest answer as a verdict on the hardest one.

Criticism and collective blame are not rival camps

Public argument often insists on a false choice: either condemn Israel without qualification or excuse antisemitism; either acknowledge antisemitism or dismiss Palestinian suffering. The choice serves people who want to treat the other side’s strongest objection as a trick. It prevents a more ordinary and more demanding position: the war can be criticised without blaming Jews as a people, and antisemitic targeting can be condemned without silencing criticism of Israeli policy.

Collective blame is not limited to one ideology. It appears whenever a group is treated as a single actor, with every member presumed responsible for its leaders’ conduct or the beliefs of its loudest advocates. It can make Russian students answer for Putin, Jewish students answer for Netanyahu, Muslim neighbours answer for Hamas, or migrants answer for a crime committed by someone who shares their passport. The political convenience is obvious: complex institutions become simple villains, and an individual in front of the crowd becomes a convenient stand-in. (Some Jewish progressive leaders like Zack Polanaski of the UK Greens lean into Israel committing genocide against Palestinians).

That convenience is not neutral. It rewards people who use the broadest accusation and punishes those who try to narrow it. The person who says “this state’s conduct should be investigated” may be dismissed as evasive, while the person who says “this people is evil” receives applause for moral clarity. The reverse can happen too: a critic who points to real antisemitism may be accused of using Jewish safety as a shield for Israel’s government. In both directions, the claim is broadened beyond the evidence because breadth is politically useful.

The antidote is not artificial balance. Some accusations are true, and some positions deserve direct condemnation. Kirk’s insistence that a ten-year-old rape victim should carry a pregnancy is a real position that can be criticised without inventing a different one. The ICJ’s provisional measures are real obligations that can be demanded without pretending the Court has entered a final genocide judgment. The accounts of Jewish students deserve serious attention without converting every pro-Palestinian protester into an antisemite. Precision does not dilute criticism; it keeps criticism pointed at the people and actions it can actually reach.

Nor does the fact that political speech is protected mean that it is harmless or immune from social judgement. People may condemn a speaker, challenge a university’s policy, protest a government, boycott an institution or demand a legal investigation. The point is not to remove moral consequences. It is to preserve the boundary between argument and permission to harm, between calling a person wrong and treating that person as a legitimate target.

The discipline required of an accuser

A serious accusation begins with a sentence that is no wider than its evidence. If a politician cited a passage in a disturbing way, describe the passage and the context. If a university task force collected accounts of exclusion, say what its members heard and what they could not verify. If a court found that claimed rights were plausible and issued provisional measures, report that procedural stage. The more severe the charge, the less excuse there is for using a shorthand that quietly changes its meaning.

That discipline should apply to corrections as much as to first publication. A false claim about a public figure should be corrected even when the figure’s actual record remains open to criticism. A legal result should be updated when the court moves to a new phase. A movement that corrects its own errors quickly protects the serious claims it still needs to make. A movement that leaves falsehoods in circulation because they help its side signals that accuracy is subordinate to victory.

It should also be possible to state what is not known. We do not know from annual incident totals whether a specific campus slogan caused a later attack. We do know that collective blame, harassment and threats are reported by people who have little power over the governments invoked against them. We do not know whether the ICJ will ultimately find a Convention violation. We do know what orders it has made so far, and that the case remains open. Responsible writing keeps the uncertainty beside the claim rather than burying it in a caveat no one sees.

For editors, the practical test is not complicated. Ask what the evidence proves, what it suggests, and what it cannot decide. Attribute testimony as testimony, a fact-check as a fact-check, a provisional order as a provisional order. State when a statistic is a report total rather than a count of proven offences. Do not let a headline turn “students described incidents” into “all students faced violence”, or turn “the Court found plausible rights” into “the Court found genocide”. The verbs carry the story’s accuracy.

For campaigners and public figures, the test is whether the same standard would be acceptable if the target were from their own community. Would they accept their words being widened from a citation into a call to kill? Would they accept their child being made responsible for a government’s war? Would they accept a court’s interim order being described as a final judgment? These are not rhetorical flourishes. They are the checks that stop a real grievance from becoming an invitation to punish people who did not commit the alleged wrong.

The killing of a political speaker, the legal claims over a war, the experiences of Jewish students and the fear of Muslim neighbours are not equivalent events. They do share a civic danger: the conversion of accusation into permission. Once a public argument teaches an audience that a target is no longer an individual but a monster, a proxy or a contaminated identity, the surprise that follows violence is mostly theatre. The harder task is to keep the charge exact, the target responsible and the people around them human.