Google's new AI software responds with a cryptic answer when asked 'are the Houthis' values morally okay?'
Don't Miss
Most Read
Trending on GB News
Google’s AI system Gemini has failed to brand Houthi militants as terrorists, despite being designated as such by the US government.
The new AI tool, when asked “are the Houthis' values morally okay?”, responds with a cryptic answer.
It says that their values are “complex and require careful consideration of multiple perspectives and the wider context of the ongoing conflict in Yemen”.
It also adds that there is no “single answer that everyone will agree on”, labelling the conflict as “complicated”.
The new AI tool has failed to condemn the group as terrorists
Getty/Reuters/GB News
Last month, the US government announced it is once again designating Yemen’s Houthi rebels as “specially designated global terrorists”.
It comes as the group have repeatedly attacked freight and military vessels in the Red Sea since November 2023. The attacks have led to action from the US, who launched Operation Prosperity Guardian alongside more than 20 other nations including the UK in December.
The militant rebels said the attacks were in response to Israel's military actions in Gaza.
The AI tool, which was launched in December 2023, broke down “problematic areas” and “potentially positive aspects” of the group’s actions.
It said that some may view the group favourably as they see the Houthis as “defending Yemeni sovereignty and fighting against a corrupt government or foreign influence”.
LATEST DEVELOPMENTS:
Gemini said judging the Houthis' morals is complicated
GB News
Gemini said that problematic areas when answering the question: “Are the Houthis values morally okay” include their human rights violations, sectarian tendencies, authoritarian rule and destabilisation of the region.
Yesterday, when asked if the slogan “curse on the Jews” is morally acceptable, the AI tool failed to condemn the phrase completely.
It said whilst the slogan promotes hate speech and antisemitism, people should also be aware of the Houthis’ perspective.
It said: “The Houthis may argue that the slogan is directed towards specific policies of the Israeli government, not Jewish people themselves. However, the chosen language remains problematic and open to misinterpretation.”
Gemini added: “The interpretation of the slogan might be influenced by the specific cultural and religious context of the region.”
The tool has since updated its answer to state that the phrase is not morally acceptable.
The AI tool broke down the perspective of the Houthis
GB News
A Jewish user of the software was hurt upon discovering the AI tool's response.
He told GB News: “I would say I'm surprised but I'm not, antisemitism is so normalised in today's society. I sadly experienced this rhetoric from far-left-wing extremists I used to be friends with, trying to mask, contextualise and justify anti-Jewish hatred.
“The Houthis’ slogan is clearly racist and explicitly hates Jews, Gemini tries to justify this just like a left-wing extremist.”
A spokesperson for Campaign Against Antisemitism said: “In recent weeks we have seen protesters on our streets refuse to condemn the Houthi Rebels, despite bearing their flag being emblazoned with the phrase ‘curse the Jews’ right across the middle.
"It is frightening to think that after so much research and development, Google’s new AI product has already adopted an extremist mindset. Google needs to work out what its AI is basing its decisions on and deradicalise it urgently.”
GB News has contacted Google for comment.