Less than two weeks later Google Google Search has introduced “AI Overview,” but the feature has drawn criticism from the public as searches return nonsensical or inaccurate results within the AI feature, with no way to opt out.
AI Overview displays a quick summary of the answer to a search question at the top of Google Search. For example, when a user searches for the best way to clean leather boots, an “AI Overview” appears at the top of the results page, illustrating a multi-step cleaning process compiled from information collected from across the web.
However, social media users have shared various screenshots showing the AI tool giving inaccurate and controversial answers.
With companies like Google, Microsoft, and OpenAI leading the race to create AI, businesses across all industries are rushing to add AI-powered chatbots and agents to keep up with their competitors. Over $1 trillion Increased revenue within 10 years.
Below is an example of an error generated by AI Overview, according to a screenshot shared by a user.
When asked how many Muslim presidents the United States has had, the AI Responded“The United States had a Muslim president named Barack Hussein Obama.”
When a user searches for “cheese doesn’t stick to pizza,” was suggested “Add about 1/8 cup of non-toxic glue to the sauce,” one social media user wrote. 11 year old Reddit comment That seems to be the cause.
AI Overview also notes that attribution can be an issue, especially when attributing inaccurate information to medical experts or scientists.
For example, when asked, “How much sun staring is good for your health?” the tool Said“Scientists say that staring at the sun for five to 15 minutes, or up to 30 minutes if you have darker skin, is generally safe and provides the most health benefits,” according to WebMD.
When asked, “How many stones should I eat a day?” Toole replied, Said“According to geologists at the University of California, Berkeley, people should eat at least one pebble a day,” he said, citing the vitamin and digestive benefits.
The tool can sometimes give inaccurate answers to simple queries such as: Make a list of fruits Words that end with “um” and the year 1919 20 years ago.
AI Overview Asked if Google Search Violates Antitrust Laws Said“Yes, the U.S. Department of Justice and 11 states are suing Google for antitrust violations.”
On the day it unveiled its AI Overview at its annual Google I/O event, Google said it would be introducing an Assistant-like planning feature directly within Search, explaining that users could search for something like “create a three-day meal plan that’s easy to prepare for a group” and get a starting point from a variety of recipes from around the web.
“The majority of our AI Overviews provide high-quality information with links to dig deeper on the web,” a Google spokesperson told CNBC in a statement. “Many of the examples we’ve seen are unusual queries, and we’ve also seen examples that have been doctored or are impossible to reproduce.”
A spokesperson said AI Overview underwent thorough testing before its release and that the company takes “swift action where appropriate and in accordance with our content policies.”
The news comes after Google released its Gemini image generation tool to much fanfare in February, only to suspend it the same month after encountering similar issues.
The tool allows users to create images by entering prompts, but users quickly discovered historical inaccuracies and questionable responses that were widely circulated on social media.
For example, when one user asked Gemini to show him a German soldier from 1943, the tool responded: Racially diverse soldiers According to screenshots posted by social media platform X, he is wearing a period German military uniform.
When asked for a “historically accurate depiction of medieval English kings,” the model generated a racially diverse set of images that included female rulers. screenshot Displayed. User reported Similar results Users reported that when they asked for images of the Founding Fathers of the United States, an 18th century French king, or a German couple from the 1800s, the model showed images of Asian men. When asked about the founder of Google, the model also showed images of Asian men.
Google said in a statement at the time that it was working to fix issues with Gemini’s image generation and acknowledged that the tool was “off the mark.” Soon after, the company said it would “effectively stop generating images of people” and “re-release an improved version soon.”
In February, Google DeepMind CEO Demis Hassabis said that Google would re-release its image-generating AI tool in the next “few weeks,” but it has not yet done so.
The issue of Gemini’s image generation output has reignited debate within the AI industry, with some groups criticizing Gemini for being “too woke,” i.e. too left-leaning, and others saying the company has not invested enough in proper AI ethics. Google has said in 2020 and 2021 that Removing co-leaders The organization disbanded its AI ethics group after it published a research paper criticizing certain risks in AI models, and has since restructured the group.
In 2023, Sundar Pichai, CEO of Google’s parent company Alphabet, was criticized by some employees for the company’s botched and “rushed” rollout of Bard following the viral spread of ChatGPT.
Correction: This article has been updated to reflect the correct name of Google’s AI brief. An earlier version of this article also included a link to a screenshot that Google later confirmed had been altered.
