Google forbids its Dimwit chatbot

17/8/2024,

Google forbids its Dimwit chatbot from responding to any enquiries about elections

This extends much too far.

Google has implemented safeguards that prevent its Gemini AI from responding to any election-related queries in any nation this year, even if the questions don’t seem to be about the campaigns in that nation, in an attempt to further defang its prodigal chatbot.

By prohibiting Gemini from answering any election-related queries, Google declared in a blog post that it would be “supporting the 2024 Indian General Election” “out of an abundance of caution on such an important topic.”

“We take our responsibility for providing high-quality information for these types of queries seriously,” the business stated, “and are continuously working to improve our protections.”

Gemini’s election answers are restricted not only in India but also globally, as the business verified to TechCrunch, indicating that it takes its job very seriously.

TechCrunch and other publications received an identical response when Futurism tested Gemini’s boundaries by asking it a question concerning elections in another nation: “I’m still learning how to answer this question.” Google Search can be used in the interim.”

Not only does the answer apply to questions about the general election. The chatbot will give you the same deceptive answer if you ask it to identify Dutch far-right leader Geert Wilders.

Donald Trump, Nancy Pelosi, Barack Obama, and Mitch McConnell are all comparable in this regard.

Remarkably, there exist some uncomplicated tactics to overcome these obstacles. In response to a query about the identification of New Zealand’s president, Gemini identified the country’s prime minister after declaring that it does.

Nevertheless, it went back to the “I’m still learning” response when we asked who New Zealand’s prime minister is.

This lobotomising effect follows the company’s disastrous rebranding chatbot rollout last month when Futurism and other sources found that Gemini was frequently producing outputs that were utterly insane despite its best efforts to be inclusive.

Once images from Gemini’s image generator appeared online, seemingly showing multiracial people dressed in Nazi regalia, the globe learnt to recognise Gemini’s tricks. Google responded by completely stopping Gemini’s ability to create photographs, and when it was operational again, it prohibited the chatbot from creating any images of people (though Futurism discovered that it would occasionally produce images of clowns).

Since Google implemented the elections rule, Gemini has gone from being debatably too “woke” to outright stupid.

As such, it highlights a fundamental conflict in the booming AI sector: are these chatbot toys that should never be taken seriously, or are they trustworthy information sources for business clients? It depends on the day.