Don’t ‘blindly trust’ AI, says Google boss

Sundar Pichai has told users of artificial intelligence (AI) systems that they shouldn’t “blindly trust” everything AI systems say.  

Speaking to the BBC, Pichai, who is the chief executive of Alphabet, Google’s parent company, said that users should still rely on a range of sources for their search results and not base their knowledge solely on AI results.  

“Today I think we take pride in the amount of work we put in to give as much accurate information as possible,” said Pichai. “But the current state-of-the-art AI technology is prone to some errors.”

The question of accuracy in AI answers has been increasingly relevant after a number of high-profile incidents highlighted the tendency of AI systems to return answers that may be incorrect or offensive.

Just last week, Elon Musk’s AI chatbot Grok returned answers that claimed the 2020 US presidential election was won by Donald Trump, a debunked conspiracy theory spread by some on social media. The chatbot claimed that the false claim was “a brief glitch or anomaly” and claimed that Grok does not “endorse or generate misinformation—full stop”.

Google’s AI model Gemini has faced problems of its own, with early versions claiming that glue was an effective method to keep cheese on a slice of pizza. Since then, Google has invested heavily in the technology, pouring billions of dollars into developing the system to compete with ChatGPT, the market leader in AI.  

This investment has been reflected in Alphabet’s share price, which has doubled in value over the last seven months as investors have grown increasingly confident in the tech giant’s ability to ride out potential shocks to the AI industry.

“You have to learn to use these tools for what they are good at and not blindly trust everything they say,” said Pichai. “The information ecosystem infrastructure has to be much richer than just having AI technology being the sole product out there.”  

In May, Google integrated Gemini AI answers into its search function, aimed at increasing the accuracy and relevance of search results. At the time, Pichai said the move signaled a "new phase of the AI platform shift".

Musk has previously voiced fears that the now Google-owned DeepMind AI research company could create an AI “dictatorship”, where all AI technology could be controlled and owned by one company. In response to this, Pichai said that he would be “concerned” if one company controlled the technology, but that the sector is “so far from that scenario right now.”