“Barack Obama is the first Muslim President of the US.” “Eat at least one small rock a day.” False, hallucinated statements like these are turning out to be a disaster for Google, its AI prowess, and CEO Sundar Pichai.
Google’s AI-powered search feature, designed to provide users with quick and concise answers to their queries, is being criticized for frequently delivering inaccurate information.
The AI Overview tool has been called out on social media platforms for its tendency to present confidently stated but incorrect responses to user questions.
While Google’s AI search aims to streamline the process of accessing information online, its recent track record has raised concerns about the reliability of its outputs and the potential consequences of spreading misinformation.
From hilarious to downright dangerous
One of the most notable examples of misinformation from Google’s AI is its promotion of the idea that engaging in fellatio could “boost your immune system.” This advice was allegedly supported by a study from the University of California, Santa Barbara, but was sourced from a satirical article in the university’s student paper.
Additionally, the claim that people engaged in such an act were “33 percent less likely to catch airborne illnesses” is a fact that it entirely made up on its own.
Misleading suggestions from the AI include cooking tips such as using petrol or gasoline in a spicy spaghetti dish to make it taste spicy and adding glue to pizza sauce to make it more sticky.
These recommendations have been traced back to humorous internet posts and a lack of credibility. While they may seem amusing at first glance, the dissemination of such misinformation can erode trust in the reliability of online information sources and lead to unintended consequences if individuals attempt to follow the advice provided.
Beyond the humorous inaccuracies, Google’s AI has also provided severe misinformation on religion and politics. For example, it falsely stated that Barack Obama was the first Muslim president of the United States and incorrectly claimed that President Joe Biden has no formal religious affiliation.
These errors not only misrepresent historical facts but also have the potential to perpetuate harmful stereotypes and misinformation about public figures.
What exactly is Google’s accountability?
In response to the criticism, Google acknowledged the errors but stated that they occur in uncommon queries and do not reflect most users’ experiences.
However, many argue that even rare instances of misinformation are unacceptable, especially considering Google’s dominant position in the search market.
As one of the primary gateways to online information, Google is responsible for ensuring the accuracy and reliability of the content it presents to users.
Google CEO Sundar Pichai has described the inaccuracies as an “inherent feature” of AI large language models (LLMs) and admitted that they are still an unsolved problem.
A problem complex to solve
Despite efforts to improve the accuracy of AI Overview, Pichai acknowledges that errors will continue to occur as AI systems grapple with the complexities of natural language processing and understanding context.
The development of AI technology presents numerous challenges, particularly when it comes to interpreting and generating human language. While AI systems have made significant strides in recent years, they are still prone to errors and misunderstandings, particularly in nuanced or ambiguous situations.
As Google continues to refine its AI algorithms, it must balance the pursuit of innovation with the need to maintain trust and accountability in its products.
While Pichai remains optimistic about the usefulness of AI Overview, the online backlash against the feature continues to grow. Critics argue that Google should focus on developing AI for more valuable applications beyond search, rather than competing with other AI models.
While AI has the potential to revolutionise various industries and improve the efficiency of everyday tasks, its current limitations highlight the need for cautious optimism and ongoing research to address its shortcomings.
