Hackers duped Meta AI support chatbot to steal celebrity Instagram accounts

Cybercriminals tricked Meta's AI customer support chatbot into handing over control of high-value celebrity Instagram accounts.

In breve

The provided raw text from Middle East Eye is a news article about the UK government denying entry to American political commentators Hasan Piker and Cenk Uygur. The stated reason from the Home Office is that their presence 'may not be conducive to the public good,' but the specific comments or actions were not specified. Piker and Uygur claim the ban is due to their criticism of the Israeli government. The article covers social media reactions, including support for the commentators and criticism of the UK government for perceived hypocrisy. This article is entirely unrelated to the provided topic, which is about hackers duping a Meta AI support chatbot.

Punti chiave

• American political commentators Cenk Uygur and Hasan Piker were denied entry to the UK by the UK government.
• The UK Home Office cancelled their Electronic Travel Authorisations (ETAs) on the grounds that their presence in the UK may not be conducive to the public good.
• Cenk Uygur and Hasan Piker claim the denial was motivated by their criticism of the Israeli government.
• The Home Office did not specify which comments or actions formed the basis for the decision.
• Social media users contrasted the ban with the UK allowing entry to Israeli President Isaac Herzog, who has been accused by the UN of statements amounting to incitement to commit genocide.

Contesto

A group of hackers successfully exploited Meta's artificial intelligence-powered customer support chatbot to steal and resell premium Instagram account handles, including those belonging to celebrities and brands, before the company patched the vulnerability, according to sources familiar with the incident. The attackers manipulated the AI chatbot by posing as legitimate account owners and submitting fraudulent support requests, tricking the system into resetting login credentials and transferring ownership of sought-after usernames. The scheme targeted so-called "OG handles" — short, memorable, or single-word usernames that can fetch thousands of dollars on the black market due to their scarcity and prestige. The breach, which remained active for an undisclosed period, highlights the growing risk of AI-driven customer service systems being weaponized by cybercriminals. Security experts…

Lettura DEO

Verdetto: REJECT - Topic mismatch: The article content does not correspond to the specified news event.
Confidenza: 15/100

The input topic describes a specific news event involving hackers exploiting Meta's AI support chatbot to steal celebrity Instagram accounts. However, the provided article preview and all structured data refer exclusively to a separate, unrelated event: the UK denying entry to political commentators Hasan Piker and Cenk Uygur. The structured data itself notes this mismatch (claim index 9). The article is real and verifiable for the UK travel ban story, but it does not address the assigned topic at all. Publishing this as a story about the Meta AI hack would be factually incorrect and misleading. Therefore, the article cannot be published under the given topic. Libre judge fallback via DeepSeek Gamma.

Cosa resta incerto

• Critical mismatch between input topic and article content: input topic is 'Hackers duped Meta AI support chatbot to steal celebrity Instagram accounts', but the provided article is about Hasan Piker and Cenk Uygur being denied entry to the UK.
• The structured data extracts claims and evidence from the article, but the article itself is about a completely different event than the one specified in the input topic.
• No evidence in the provided article supports the existence of a Meta AI hack, a celebrity Instagram account theft, or any related incident.

Categoria: cronaca
Entità: Hackers, Meta, Instagram