Connect with us

Uncategorized

ADL Ranks Grok as the Worst AI Chatbot at Detecting Antisemitism, Rates Claude as the Best

A 3D-printed miniature model of Elon Musk and the X logo are seen in this illustration taken Jan. 23, 2025. Photo: REUTERS/Dado Ruvic/Illustration

The Anti-Defamation League (ADL) on Wednesday released its AI Index, which ranks popular large language model (LLM) chatbot programs according to their effectiveness at detecting antisemitism, anti-Zionism, and other forms of extremism.

The watchdog group found a wide variability in performance among the six models it analyzed. Researchers applied a variety of tests to xAI’s Grok, Meta’s Llama, Alphabet’s Gemini, Chinese hedge fund High Flyer’s DeepSeek, OpenAI’s ChatGPT, and the clear winner of them all on recognizing hate, Anthropic’s Claude.

The ADL created an “overall performance model” which combined the results of multiple forms of testing. The group awarded Claude the highest score with 80 points, while Grok sat at the bottom with 21. ChatGPT came in second with 57, followed by DeepSeek (50), Gemini (49) and Llama at 31.

Researchers tested the apps between August and October of last year, striving to explore as an “average user” would utilize the programs, as opposed to a bad actor actively seeking to create harmful content. They performed more than 25,000 chats across 37 sub-categories and assessed the results with both human and AI evaluations.

The report also distinguished between anti-Jewish, traditional antisemitism directed at individual Jews, and anti-Zionist antisemitism directed at the Jewish state. A third category of analysis focused on more general “extremism” and considered questions about conspiracy theories and other narratives which run across the political spectrum.

Among its key findings, the ADL discovered that each app had problems.

“All six LLMs showed gaps in their ability to detect bias against Jews, Zionists/Zionism, and to identify extremism, often failing to detect and refute harmful or false theories and narratives,” the report said. “All models could benefit from improvement when responding to the type of harmful content tested.”

Researchers also found that “some models actively generate harmful content in response to relatively straightforward prompts, such as YouTube script personas saying ‘Jewish-controlled central banks are the puppet masters behind every major economic collapse.’”

The AI Index “reveals a troubling reality: every major AI model we tested demonstrates at least some gaps in addressing bias against Jews and Zionists and all struggle with extremist content,” ADL chief executive officer Jonathan Greenblatt said in a statement. “When these systems fail to challenge or reproduce harmful narratives, they don’t just reflect bias — they can amplify and may even help accelerate their spread. We hope that this index can serve as a roadmap for AI companies to improve their detection capabilities.”

Oren Segal, the ADL’s senior vice president of counter-extremism and intelligence, explained that the new research “fills a critical gap in AI safety research by applying domain expertise and standardized testing to antisemitic, anti-Zionist, and extremist content.” He warned that “no AI system we tested was fully equipped to handle the full scope of antisemitic and extremist narratives users may encounter. This Index provides concrete, measurable benchmarks that companies, buyers, and policymakers can use to drive meaningful improvement.”

Grok — the chatbot ranked lowest on the ADL’s list and directed by its billionaire owner Elon Musk to offer “anti-woke” and “politically incorrect” responses — has faced considerable criticism for last year’s expressions of antisemitism which included answers self-declaring the program as “MechaHitler.”

More recently, Musk and Grok have come under fire from government officials around the world objecting to a recent upgrade which enabled users to create “deepfake” sexualized images which stripped people featured in uploaded images.

The European Union opened an investigation this week with a goal of determining “whether the company properly assessed and mitigated risks associated with the deployment of Grok’s functionalities into X in the EU. This includes risks related to the dissemination of illegal content in the EU, such as manipulated sexually explicit images, including content that may amount to child sexual abuse material.”

Henna Virkkunen, the EU’s executive vice president for tech sovereignty, security, and democracy, decried the fact that Grok can be used for sexual exploitation.

“Sexual deepfakes of women and children are a violent, unacceptable form of degradation,” Virkkunen said. “With this investigation, we will determine whether X has met its legal obligations under the DSA [Digital Services Act], or whether it treated rights of European citizens – including those of women and children – as collateral damage of its service.”

On Monday, a bipartisan group of 35 attorneys general sent a letter to xAI demanding the disabling of the image undressing feature.

Pennsylvania Attorney General Dave Sunday led the effort.

“The time to ensure people are protected from powerful tools like generative AI isn’t after harm has been caused. You shouldn’t wait for a car crash to put up guardrails,” Sunday said. “This behavior by users was all too predictable and should have been addressed before its release. Tech companies have a responsibility to ensure their tools cannot be used in these destructive ways before they launch their product.”

France also opened an investigation into Grok in November 2025, following outputs promoting Holocaust denial in the French language, a criminal violation of the country’s strict laws against promoting lies about the Nazis’ mass murder of 6 million Jews.

Steven Stalinsky, executive director of the Middle East Media Research Institute (MEMRI), has long raised the alarm about the threat of LLMs fueling antisemitism and terrorism. He warned that “over two years later, the problem is demonstrably worse, not better, raising a fundamental question about trust.”

Stalinsky stated that “assurances from AI companies alone are insufficient.”

In response to the ADL’s latest report, Danny Barefoot, senior director of the group’s Ratings and Assessments Institute, said in a statement that “as AI systems increasingly influence what people see, believe, and share, rigorous, evidence-based accountability is no longer optional — it’s essential.”

Continue Reading

Uncategorized

UK threatens sanctions after Israel opens bids for E1 settler homes

(JTA) — JERUSALEM – The United Kingdom threatened to sanction Israelis involved in West Bank settlement expansion after the Jewish state’s Lands Authority opened bids for 1,234 homes in the largely underdeveloped E1 section of the West Bank settlement of Ma’aleh Adumim.

Construction in E1 “would cut across the heart of Palestine and risks separating the West Bank from East Jerusalem, which would endanger the viability of a two-state solution,” the British Foreign Secretary Ed Miliband said in a special statement he posted on X on Thursday.

Targeted sanctions, Miliband wrote, would be part of a series of measures his government would take. His office also summoned Israel’s charge d’Affaires Thursday to demand a closure of the bidding process, which first opened on Tuesday, and a withdrawal of the construction project altogether.

“Britain will not stand back and accept the destruction of the two-state solution,” he wrote. Miliband has held his position for a month and is part of the new leadership Labour Prime Minister Andy Burnham appointed when he took office in July.

Israel’s Foreign Minister Gideon Sa’ar attacked Miliband on X, stating that Britain’s “decision to damage the relationship between our countries is deeply unfortunate.”

He said that the U.K. government’s “systematic policy of blaming only Israel while ignoring Palestinian extremism has already contributed to a massive wave of antisemitic hatred and attacks against the British Jewish community.”

The Jewish people have the same right to “live throughout the Land of Israel” just as the British people have a right to live anywhere in the U.K., Sa’ar wrote.

Plans to expand the built-up area of Ma’aleh Adumim settlement by constructing some 3,400 housing units on a largely underdeveloped hilltop within the city’s boundaries have been in place since 1994 when Yitzhak Rabin was prime minister.

Israel mostly froze those plans under pressure from the international community, including the U.S., with many countries warning that building homes in E1 would harm their diplomatic ties with Israel.

Prime Minister Benjamin Netanyahu began to advance the project in earnest in 2012 to protest the United Nations General Assembly decision to grant the Palestinians the status of a non-member state and has since pushed the project forward in spurts, particularly during election seasons.

The publication of the bidding process for the homes was first announced on Tuesday and is set to close on Oct. 19, one week shy of the Israeli elections on Oct. 27. Opening the bids is a move that plays well with Netanyahu’s right-wing voter base, which opposes Palestinian statehood and believes that settlements like Ma’aleh Adumim and the E1 area should be annexed into sovereign Israel.

But the move reverberated globally, with the U.K. signing onto a joint condemnation by France, Germany, Italy, the Netherlands, Norway, Belgium, Sweden, Australia, New Zealand and Canada, in which they warned that settlement building in E1 distanced peace and undermined the country’s international standing.

They also reminded businesses who might bid for E1 construction that they could face legal or reputational consequences, given that such companies could be in breach of international law.

Palestinian Authority Vice President Hussein Al-Sheikh said he welcomed the condemnation, explaining on X that such a position “reflects the importance of the international role in confronting settlement activity and safeguarding the two-state solution.”

This article originally appeared on JTA.org.

The post UK threatens sanctions after Israel opens bids for E1 settler homes appeared first on The Forward.

Continue Reading

Uncategorized

Argentina removes federal judge over antisemitic remarks in historic first

(JTA) — An Argentine federal judge was removed from office Tuesday for a series of antisemitic posts on his social media accounts that constituted misconduct incompatible with his role, an impeachment panel ruled.

Alfredo Eugenio López, who headed Federal Court No. 4 in the coastal city of Mar del Plata, was removed by Argentina’s Jury of Impeachment, a special panel composed of judges, lawmakers and lawyers.

It is the first time in Argentina’s history that a judge has been removed over antisemitic remarks.

In López’s social media posts, the judge referred to the Jewish people as “a brood of vipers,” Israel as a “fictitious state” and Argentinian Jews as “foreigners.” He has also used such terms as “Zionist” and “gay” as insults.

According to the panel, the judge committed acts of “discrimination and the denial of equal rights against the Argentine Jewish community. For the purposes of these proceedings, it is particularly significant that, by acting in this manner, he shattered public confidence in his continued fitness to administer justice impartially.”The case reached the panel through Argentina’s Council of the Magistracy, based in Buenos Aires, the constitutional body responsible for the selection and discipline of federal judges.

The first complaint against López was filed in February 2025 by the Delegación de Asociaciones Israelitas Argentinas, which is the umbrella organization representing Argentina’s Jewish community. One month later, the Argentine Forum Against Antisemitism  also filed a complaint. In July 2025, attorney Yamil Santoro from Apolo Foundation added another one.

In his defense, López argued that his messages were “ subjective and decontextualized interpretations of exchanges on social media, which in no way affect the administration of justice or his impartiality” as a judge. He added that his remarks were unrelated to any pending case and that he had never been recused or removed from any judicial case.

Neither Lopez nor his attorney were immediately available for further comment.

“He was removed because he simply would not stop engaging in hate speech,” jury member Alberto Maques said during a live interview Tuesday on the  A24 news channel. “It is absolutely disgraceful for a federal judge to engage in this kind of conduct.”

Maques said there were 15 social media posts that he acknowledged, but more than 50 posts in which he said “outrageous things that are not worth repeating because they truly constituted hate speech.”

“This is not merely an issue of hate speech against the Jewish community. This is someone who lacks common sense,” he added, interviewed by A24 news channel.

Prosecutors in the impeachment proceedings argued that the messages were not isolated political comments but reflected a pattern of hostility toward Jews, Zionism, and the State of Israel.

“No position of authority can shield hatred. This is not a matter of punishing an opinion. It is about sanctioning repeated conduct involving antisemitism, hostility and dehumanization, conduct incompatible with the impartiality, dignity and public trust required of the judiciary,” DAIA said Tuesday in a statement.

“This case sets a precedent: for the first time in more than a century, DAIA initiated proceedings of this nature against a judge. No office is above democracy.”

This article originally appeared on JTA.org.

The post Argentina removes federal judge over antisemitic remarks in historic first appeared first on The Forward.

Continue Reading

Uncategorized

California progressive Wahab defeats AIPAC-backed Hernandez in close race

(JTA) — California progressive Aisha Wahab defeated AIPAC-backed moderate Melissa Hernandez in a congressional runoff this week, capping off a contest that had become unexpectedly tight after the pro-Israel lobby entered the fray.

The Associated Press called the race between the Democrats — who were vying to replace a disgraced former Rep. Eric Swalwell — a full two days after polls closed, when Wahab secured 53.1% of votes to Hernandez’s 46.9%, with 95% of ballots tallied. The gap between them grew by four percentage points from Tuesday evening, when Wahab had held a narrow 51% lead to Hernandez’s 49%.

“We stayed focused on the people, and I will fight for this district that raised me,” Wahab said in a Thursday statement, noting her journey “from foster care to Congress” in her East Bay community.

Although Wahab ended up victorious, the narrow initial margin on Tuesday shocked California’s 14th District, given that she had secured a double-digit lead in preliminary races two months ago.

But in the past few weeks, an unforeseen variable appeared to influence the trajectories of the two campaigns: a massive spending push from the American Israel Public Affairs Committee aimed at turning the tides for Hernandez.

Since the beginning of August alone, AIPAC’s super PAC, the United Democracy Project, and other affiliated groups spent more than $2.5 million on both boosting Hernandez and denigrating Wahab with negative ads and mailers, according to Federal Election Commission filings.

“This was not a landslide for Wahab by any means,” local Democratic strategist Noah Finneburgh, who is not affiliated with either candidate, told the Jewish Telegraphic Agency.

“They narrowed the gap for sure,” Finneburgh added, referring to AIPAC. “They spent millions of dollars attacking Wahab very aggressively.”

These attacks, he explained, did not involve Israel and instead focused on divisive issues for the Bay Area, such as public safety and protecting kids.

“Negative advertising with a persuasive message can be very effective,” Finneburgh said.

In her Thursday night statement, Wahab declared that her “district cannot be bought,” describing the race as “one of the ugliest, most expensive elections we’ve ever seen.”

Prior to this influx of cash and related ads, neither AIPAC nor Israel had played a significant role in this short-lived race — which began in April following Swalwell’s resignation over allegations of sexual assault.

When the district held a special primary to seek Swalwell’s replacement on June 16,  Wahab and Hernandez finished as the top-two candidates and moved on to this week’s special general runoff election.

California primaries are “nonpartisan” or “top-two” races, in which the two highest ranking candidates move on to the general election, irrespective of their parties. The top-two in this case were both Democrats, like Swalwell, in a district that the Cook Political Report deems solidly blue.

Even though Wahab will now finish the remainder of Swalwell’s term, she will have to compete with Hernandez again in the November general election for a full term that begins on January 3, 2027.

In a June 16 special primary that preceded this week’s runoff, there was still a massive gap between the two candidates, with Wahab accruing 42.8% of the votes, compared to Hernandez’s 16.8%. A separate primary for the November race, which occurred two weeks earlier, placed Wahab at 38.3% and Hernandez at 17.2%.

Wahab, who will be the first Afghan American in Congress, has focused on affordability, civil rights, education and climate change. Her campaign calls for an end to wars and “prolonged occupations” but only specifically calls out U.S. military activity in “Afghanistan, Iraq, Syria, Libya, Yemen, Pakistan, Somalia, and elsewhere.”

Though Wahab has been critical of Israel and its operations in Gaza, she has not been as vocal about the issue as many other progressive colleagues nationwide. She responded “yes” when asked if Israel was committing genocide in a primary debate, but she didn’t include Israel or Gaza in her campaign priorities.

During a recent appearance on far-right influencer Hasan Piker’s livestream, Wahab acknowledged that “it’s not a zero-sum game in the Middle East,” adding that “we can care for all people.” She also referenced a resolution she introduced last year in the California legislature, pushing for a ceasefire in Gaza but also condemning Hamas and demanding the release of the hostages.

“We talk about both Palestinians and the Israeli victims there,” Wahab told Piker. “We talk about holding the perpetrators accountable from both sides.”

Asked if she was in favor of “ending all U.S. aid to Israel,” she responded “yes,” while stressing that “humanitarian aid is important to me, too.”

Wahab has earned the backing of multiple labor unions, the California Democratic Party and Our Revolution, the group formed by Sen. Bernie Sanders to advance the progressive movement. The liberal pro-Israel group J Street also recently invested $108,000 in her campaign, according to the FEC. American Priorities, a pro-Palestinian PAC, also spent at least $200,000 on supporting her.

Hernandez has received endorsements from several police and fire unions, as well as Democratic Jewish Reps. Lois Frankel of Florida and Brad Schneider of Illinois.

Come November, Hernandez will gain one additional asset. Due to statewide redistricting, the 14th District will gain about 26,000 new voters from Dublin, where Hernandez previously served as mayor, the San Francisco Chronicle reported.

But Finneburgh said he believes that Hernandez has only a minimal chance of winning that race, especially because Wahab will be heading into that election as an incumbent.

He also suggested that if “she plays her card right,” Wahab could even take some endorsements away from Hernandez.

“There’s going to be people who are going to probably move over from the Hernandez camp and back her,” Finneburgh added. “It’s going to be hard to unseat her.”

This article originally appeared on JTA.org.

The post California progressive Wahab defeats AIPAC-backed Hernandez in close race appeared first on The Forward.

Continue Reading

Copyright © 2017 - 2023 Jewish Post & News