The AI gravy train rolls on unchecked and the latest body on the tracks is Google (NASDAQ:GOOGL).
The tech giant’s latest experimental search feature, AI Overviews, has this week faced backlash for producing erroneous and potentially harmful information.
Rolled out on Chrome, Firefox and the Google app browser, AI Overviews uses generative AI technology, similar to ChatGPT, to summarise search results, saving users from clicking through links and doing their own research.
But some of these AI-generated summaries have proven to be disastrously inaccurate.
Cats on the moon
While the feature can correctly advise users to store bananas in a cool, dark place away from apples to keep them fresh, it has also suggested that "astronauts have met cats on the moon".
It has also ventured into dietary advice, such as recommending the consumption of "one small rock per day" for minerals and vitamins and telling users to put glue in their pizza.
These bizarre recommendations shine a light on a pivotal issue with generative AI: it doesn't discern truth from popularity.
The AI algorithms prioritise widely read content, even if it's satirical or false, over verified factual information. They’re also gullible, deriving their ‘facts’ from satirical sites like The Onion – a fraught activity even for a human of reasonable IQ.
Google's reliance on AI-generated data without adequate oversight raises concerns about the reliability of its search results and stands to erode trust in its ability to provide accurate information.
Bad for the brand
Considering the enduring verb the global megatech has spawned – ‘Google it, mate’ – this isn’t great for its brand.
The company risks undermining its lucrative search business model, which relies on users clicking on ads and links.
If users simply read AI-generated summaries, Google's revenue from ad clicks could decline.
More importantly, the issue underscores the broader societal risks of integrating AI tools without robust safeguards to filter out the misinformation and biases inherent in web content.
As the AI race intensifies, critics are calling for time out to put in some standards and guardrails, comparing the issue to pharmaceutical companies being made to ensure their products are safe before release.
Google CEO Sundar Pichai has previously said he backs a cautious and responsible approach to AI deployment but the urgency to keep pace with competitors like OpenAI and Microsoft (NASDAQ:MSFT) appears to have muted those concerns for now.