Google's AI portrays Obama as a Muslim leader.
Google's new AI overviews in the USA are causing quite a stir, with some humorous and even concerning results. The software, designed to provide direct answers to users' questions, is making mistakes that range from recommending gluing cheese onto pizzas to claiming Barack Obama was the first Muslim president.
Since AI overviews were added to the search engine, users have noticed glaring errors in the software's responses. One example is advising to attach cheese to a pizza using non-toxic glue. Additionally, the AI answered that dogs participated in the NBA and NFL leagues, further displaying its inability to differentiate between jokes and serious information.
The main purpose of these AI summaries, also known as 'AI Overviews,' is to provide users with a concise response rather than directing them to a list of links. Several startups hope to employ this strategy to compete with Google's dominance in web searches. Google, however, plans to expand this feature globally by year-end.
Some website operators and media outlets are concerned that these AI-generated summaries will reroute users away from their sites, negatively impacting their businesses. Google, on the other hand, claims that there is increased traffic to the sources of information contained within the summaries. Regardless, the long-term effects on other entities are uncertain.
AI's confusion between humor and fact
An unforeseen issue with the widespread integration of AI overviews is the software's lack of discernment between humorous statements and genuine information. For instance, the AI has erroneously cited sources for seemingly absurd claims, such as geologists recommending eating a small stone every day.
According to a Google spokesperson, the mistakes are primarily attributed to "very unusual requests" and aren't representative of the average user's experience. The "isolated examples" will be used to enhance the AI's performance. Previously, Google had to face mockery online due to their AI program, Gemini, which generated pictures of non-white Nazi soldiers and settlers. They attributed this blunder to not programming exceptions for scenarios where diversity was clearly out of place. In response, the Gemini Group stopped creating images of people altogether. (Images of stones, on the other hand, seem safe for the time being.)
In February, Google faced public shots for Gemini's missteps, gaining much attention and criticism. This reveals one of the challenges tech companies face when relying heavily on AI: ensuring it can comprehend the nuances between humor, satire, and actual facts.
Read also:
- Lack of snow also opens up new opportunities for winter tourism
- Abrupt end to e-car subsidies
- The chemical industry has little confidence
- Intersport boss hopes for sales boom through sporting events
The digital economy is greatly influenced by the advances in artificial intelligence, as seen with Google's AI summaries causing both comedic and concerning errors. Moreover, the concerns about AI's inability to differentiate humor from fact extend beyond Google's AI Overviews, affecting the entire digital economy.
Source: www.ntv.de
