Earlier this month, Google released a new artificial intelligence feature in its search engine that is aimed at helping users get relevant information from the internet more efficiently. But shortly after the tool was released, it started producing false data and Google had to remove some of the search results temporarily.
Among the main topics that occurred was the false statement regarding the former president of the United States, Barack Obama. The AI summary is wrong on the claim that Obama is a Muslim when, in fact, he is a Christian. Another of the mistakes stated that every country in Africa begins with K leaving out Kenya from the list of 54 countries. Google said that those incorrect summaries were taken down because they were against the firm’s guidelines.
According to Colette Garcia, Google’s spokesperson, most of the time, the information provided in an AI-generated summary is high-quality, and the readers are also provided with reference links. She also said that some of the errors posted online could have been false. Google piloted the AI tool before its roll-out and is sensitive to the ability of users to shape its services. It should also be noted that the company reacts very actively to mistakes made under content policies.
Every AI summary has attached the note that the technology is experimental. Google also has trials in order to avoid the appearance of unsuitable or low-quality search results. This new AI tool is Google’s bid to implement its Gemini AI into all its products in line with competition from other tech companies like OpenAI and Meta. However, the recent mistakes prove that AI can sometimes be very sure of the erroneous data it provides, which threatens to taint Google’s reputation as a credible source of information.
Sometimes, even in relatively mundane search queries, Google’s AI assistant has given the wrong or at least unclear information. For instance, when the AI was being asked questions about the sodium content in pickle juice, it was quite equivocal. It stated that it has 342 milligrams of sodium in the 8-ounce serving and 690 milligrams in the 3-ounce serving, which is confusing. Therefore, its product label says that 1 ounce of Best Maid pickle juice contains 250 milligrams of sodium.
Drawing on the data that CNN used to train Google’s AI, the AI overview stated that it was uncertain whether Google avoids using copyrighted data in scraping data, which raises a major concern with AI systems.
This is not the first time Google has faced concerns about its AI tools. In the year, the company was compelled to pause an AI photo generator because it had accuracy issues and generated photos that depicted people of color in the place of whites.
Consumers in the regions, where the two AI search overviews are presently active can turn on/off the feature on Google’s Search Labs webpage. This way users can choose if they want to continue using the AI summaries while Google is refining and developing this technology.
Robots Reading the News? DD Kisan Debuts AI News Anchors – Will They Be a Hit?
Ever since the Chinese Explore launched AI news anchors, similarly, India is getting ready to launch AI Krish and AI Bhoomi with DD Kisan on May 26. These AI anchors will help deliver farmer-specific news in 50 regional and international Indian languages online. This step has placed DD Kisan in a special category of Indian stations being the first government-sponsored station to have AI anchors on its platform.
The Ministry of Agriculture also pointed out that these AI anchors will update useful information for farmers non-stop. Some of them will be research done in agriculture, climate prediction, markets for agriculture products, government strategies, and the like.
This is also in line with the findings from the growing popularity of Indian news channels adopting AI technology. For instance, Odisha TV has an AI announcer named Lisa; Aaj Tak has an AI newsreader named Sana; News18 Punjab Haryana has an AI news presenter named Kaur; Power TV Kannada introduced Soundrya last year.
The main advantage of using AI anchors is that the anchors can work for extended periods without boredom and can produce more work in a single day than a human anchor can do. However, there are some concerns about loss of job opportunities and other technical incidents that may demoralize the show’s audience.
Holding of AI anchors in the DD Kisan and an announcement to launch the IndiaAI mission prove the government’s seriousness about AI technology. This mission is to create a holistic ecosystem of artificial intelligence in India and has been accorded a double of 20,000 crore for the next five years.
In addition, the government of India has undertaken its own venture called the BHASHINI project to work on AI translation to provide official website access in multiple languages. With regard to the above, the Ministry of Electronics and Information Technology is currently coming up with policies to be adopted to regulate innovation but not restrict it.
AI has quite a bright future in India and the GenAI is expected to be valued at 17 billion US dollars by 2030. Due to the enhanced advancement of AI, it is supposed to be available in the agricultural and media sectors for more opportunities and potentialities.
Should You Glue Your Pizza? Google’s AI Search Goes Hilariously Wrong
This week, Google announced that it will be adding AI-generated results descriptions to the search results, which could dramatically alter how we encounter information on the internet. However, this new feature proves to be extremely controversial and offers some rather peculiar and inaccurate suggestions.
People have been posting funny and, in some cases, completely wrong outputs made by Google’s artificial intelligence on social networks. For instance, one user asked a ludicrous question, which was related to diet, and he received the reply stating: ‘You should eat at least one small rock per day, ‘Thus, the answer is based on the information provided in an article published by the reputable source called The Onion, although this is actually a satire website.
Another user posed a question regarding how cheese can best adhere to the pizza and the advice gained was to add about one-eighth cup of non-toxic glue to the sauce so that it will be tackier. As can be imagined, this advice left many of the people stunned and amused.
It also recommended drinking urine to pass the kidney stone, which can be potentially lethal, despite stating that “at least 2 quarts (2 liters) of urine every 24 hours. ” It also contained fake news about former President of the United States, Barack Obama, claiming falsely that he is a Muslim while in fact he is a practicing Christian.
Google addressed these concerns by pulling down these summaries and noting that these responses were a violation of the policies they set. They stated the majority of the AI summaries send correct information and did state that some of the nonsensical appearances might have been changed or their team could not replicate such responses again.
Although the way was paved through the extensive testing of AI summaries before deploying it, Google admitted that there are drawbacks to applying the technology. Some of it they are using to make improvements on the current setup, and there are some that they are in the process of implementing.
To a larger extent, these challenges with Google’s AI summaries point to modern AI technology’s general challenges. Even after the release of such tools as OpenAI’s ChatGPT, AI has been adopted in businesses by many companies such as Microsoft. These tools are utilized in several ways for instance when a person is summarizing the notes of a meeting they just attended or when developing electronic content. Nevertheless, they are largely affected by so-called “hallucination”, when the AI produces information that does not exist to form a complete answer.
For those who do not like these AI summaries, some techniques have been developed to enable the users to uncheck Google options that display the summaries that have been created by the AI. However, it’s not the norm and many may still experience these unusual answers from their search tool.
Breastfeeding, fantasy football, dogs in basketball, or the unknown kiltomato are vivid examples of how hard it is to implement AI into present-day applications. AI becomes more advanced every day and despite this advancement, the researcher can comfortably argue that it will take more than a few years before it can be relied upon to offer judgment like that given concerning the numbers provided.
Check who got the $$ Spotlight $$ today!
- Elon Musk’s xAI Seeks $6 Billion in Funding Backed by Andreessen Horowitz. read more
- ThinkLabs AI Raises $5M in Seed Funding. read more
- German AI defence giant Helsing eyes $4B valuation in Series C funding. read more
- Mistral AI Surges in $6 Funding Amidst Competitive Landscape. read more
- AI translation startup secures $300 million funding. read more
- Orca AI secures $23M to advance development of autonomous shipping technology. read more
- Prominent ByteDance investor preps new $300m AI-centered fund. read more
- Expedera raises $20m in funding round led by indie Semiconductor. read more
- Digital invests $10.5 Million Across four AI Projects from healthtech startups. read more
- Amii Secures $3 Million From Prairiescan To fuel AI adoption in the Prairies. read more
- Angel AI Secured $4.5 Million To Provide Age-Appropriate Internet Experiences For Children. read more
- HoneyNaps secures a $11.6 million Series B investment, becoming the No. 1 ranked A.I Sleep Technology Company. read more
