Menu

AI "Hallucinations" Without Malice: Meta AI Chatbot Denied Assassination Attempt on Trump

By
Photo: AI "Hallucinations" Without Malice: Meta AI Chatbot Denied Assassination Attempt on Trump. Source: Collage The Gaze \ by Leonid Lukashenko
Photo: AI "Hallucinations" Without Malice: Meta AI Chatbot Denied Assassination Attempt on Trump. Source: Collage The Gaze \ by Leonid Lukashenko

The Meta AI chatbot denied the fact of an assassination attempt on Donald Trump and refused to acknowledge the authenticity of the shooting photos. Meta explained this as ‘hallucinations’ of artificial intelligence.

Meta Corporation, which combines Facebook, Instagram and WhatsApp, has admitted that its artificial intelligence falsely denied the assassination attempt on US presidential candidate Donald Trump. Facebook's Vice President of Global Policy Joel Kaplan assured on Tuesday, 30 July, that there was no ‘malicious intent’ behind the mistake, which ‘we regret’.

After the attempted assassination of Donald Trump, users who asked Meta AI about the shooting received answers from the chatbot with long background information and denied that the former US president had been assassinated. In addition, the famous photo of Trump after the assassination attempt with his fist raised on Facebook and Instagram was labelled as fake.

As Kaplan explained, the chatbot processes a huge amount of data accumulated over many years. But that is why it is difficult for it to work with breaking news. AI gets confused by contradictory information about recent events and does not always successfully compare them with the past. That's why Meta's management initially banned the chatbot from talking about the Trump assassination attempt. Later, the ban was lifted, but the AI continued to remain silent for some time or claim that there was no shooting.

The flood manager explained the problems with Trump's photo, in which he poses with his fist raised in the air immediately after the assassination attempt, by explaining that some attackers had posted a fake photo on the Internet, drawing smiles on the US Secret Service officers. Meta marked this cartoon as fake. But the AI started blocking all similar images, including the original one.

Meta introduced its own AI assistant, Meta-AI, in April 2024. It is supposed to help users of Facebook, Instagram, and WhatsApp. For example, the chatbot can recommend a restaurant, concert, or organise a trip. Meta-AI is available in English in more than 10 countries, including the US, Canada, Australia, South Africa, Nigeria, and Pakistan. It is not available in Europe.



Recommended

Economics

Oil Braces for a Price Plunge

11.21.2024 15:51
Politics

The Attempt to Privatise the G20 Has Yet to Succeed

11.20.2024 15:52
Life

Jodie Foster: The Actress Worth Killing For

11.19.2024 16:15
Culture

Gladiator 2: Barefoot in Sandals

11.18.2024 16:58
Politics

Europe Prepares to Seek Common Ground with Trump

11.18.2024 10:10

Similar articles

We use cookies to personalize content and ads, to provide social media features and to analyze our traffic. We also share information about your use of our site with our social media, advertising and analytics partners who may combine it with other information that you've provided to them. Cookie Policy

Outdated Browser
Для комфортної роботи в Мережі потрібен сучасний браузер. Тут можна знайти останні версії.
Outdated Browser
Цей сайт призначений для комп'ютерів, але
ви можете вільно користуватися ним.
67.15%
людей використовує
цей браузер
Google Chrome
Доступно для
  • Windows
  • Mac OS
  • Linux
9.6%
людей використовує
цей браузер
Mozilla Firefox
Доступно для
  • Windows
  • Mac OS
  • Linux
4.5%
людей використовує
цей браузер
Microsoft Edge
Доступно для
  • Windows
  • Mac OS
3.15%
людей використовує
цей браузер
Доступно для
  • Windows
  • Mac OS
  • Linux