Tag: Google Bard

Categories
Mobile Syrup

Google’s Bard AI falls short of rivals, Pichai promises upgrades to come soon

Recent criticism about Google’s Bard sharing non-factual information and rumours around the use of ChatGPT to train Bard surfacing aren’t painting a good picture for Google, and CEO Sundar Pichai is now in damage control mode.

Pichai has promised that upgrades to Bard are coming in the near future, as stated in an interview on The New York Times’ Hard Fork podcast, (via The Verge). “Pretty soon, perhaps as this  

| Open Player in New Window

goes live, we will be upgrading Bard to some of our more capable PaLM models, which will bring more capabilities; be it in reasoning, coding, it can answer maths questions better. So you will see progress over the course of next week,” he said.

PaLM is a more recent language model that is larger in scale than Bard’s current LaMDA model, and is more capable when dealing with tasks like common-sense reasoning and coding problems

Pichai acknowledged that the way Bard is now, it is failing to match the performance of rivals such as OpenAI’s ChatGPT and Microsoft’s Bing chatbot, which runs on OpenAI’s GPT-4. He also discussed concerns regarding the fast pace of development in AI models, and the threat it can pose to society.

He also acknowledged the concerns raised in an open letter signed by Elon Musk and top AI researchers calling for a six-month pause on the development of AI systems.

Despite the hurdles, Pichai remains committed to advancing AI capabilities. He emphasized the importance of anticipating and evolving to meet the challenges that AI will present, and understands that these systems are likely to become very capable, regardless of whether they reach the level of artificial general intelligence (AGI).

Check out the full podcast here.

Source: The New York Times, Via: The Verge

Categories
Mobile Syrup

Google denies using ChatGPT data for Bard training

Google has reportedly launched a new initiative called Gemini, which aims to improve its artificial intelligence (AI) chatbots after the lacklustre debut of its latest offering, Bard.

According to The Information, the tech giant is enlisting the help of its DeepMind division to beat rival OpenAI, but the report also claimed that Google may have trained Bard using data from OpenAI’s ChatGPT, taken from a website called ShareGPT.

A former Google AI researcher, Jacob Devlin, allegedly warned against using the data, claiming it would violate OpenAI’s terms of service and produce answers that looked too similar.

Despite the allegations, Google has denied using the ChatGPT data to train Bard. A spokesperson for the company, Chris Pappas, stated that “Bard is not trained on any data from ShareGPT or ChatGPT,” in a statement given to The Verge. “Unfortunately, all I can share is our statement from yesterday,” he said.

The move to involve DeepMind in Gemini is seen as a significant step, as the AI division has been striving to become more independent from Google for years. However, the collaboration could also signal a renewed effort by Google to improve its AI chatbots and compete with OpenAI.

Source: The Information Via: The Verge

Categories
Mobile Syrup

Google’s ChatGPT competitor is already sharing false information

Google’s recently-announced ChatGPT-like AI chatbot Bard is already making factual mistakes.

On Monday, February 6th, Google announced that its ChatGPT competitor Bard would go public in the coming weeks. In a blog post by Google CEO Sundar Pichai, an example of the chatbot’s capabilities was shown.

When asked, “What new discoveries from the James Webb Space Telescope can I tell my 9-year-old about?” the chatbot shared three simple-to-understand points, one of them being that “JWST took the very first pictures of a planet outside of our own solar system.”

While we know that the JWST is the most advanced telescope to peep at the universe with and can take better photos of planets outside our solar system, saying that the JWST took the very first picture of an exoplanet isn’t factual. As stated on Nasa’s website, the first image of an exoplanet was taken back in 2004, while the JWST became operational in 2022. The first exoplanet was imaged by the VLT (Very Large Telescope), stationed in Cerro Paranal, Antofagasta, Chile.

Several astronomers and astrophysicists were quick to point out the inaccuracy of Bard, with Grant Tremblay (@astrogrant) saying, “Not to be a ~well, actually~ jerk, and I’m sure Bard will be impressive, but for the record: JWST did not take “the very first image of a planet outside our solar system.” He added, “I do love and appreciate that one of the most powerful companies on the planet is using a JWST search to advertise their LLM. Awesome! But ChatGPT etc., while spooky impressive, are often *very confidently* wrong. Will be interesting to see a future where LLMs self error check.”

In a statement given to The Verge about Bard’s inaccuracy, a Google spokesperson said, “This highlights the importance of a rigorous testing process, something that we’re kicking off this week with our Trusted Tester program. We’ll combine external feedback with our own internal testing to make sure Bard’s responses meet a high bar for quality, safety and groundedness in real-world information.”

In related news, Microsoft has partnered with OpenAI to bring ChatGPT-like functionality to the Bing search engine. Read more about it here.

Via: The Verge

Categories
Mobile Syrup

Google’s ChatGPT competitor is already sharing false information

Google’s recently-announced ChatGPT-like AI chatbot Bard is already making factual mistakes.

On Monday, February 6th, Google announced that its ChatGPT competitor Bard would go public in the coming weeks. In a blog post by Google CEO Sundar Pichai, an example of the chatbot’s capabilities was shown.

When asked, “What new discoveries from the James Webb Space Telescope can I tell my 9-year-old about?” the chatbot shared three simple-to-understand points, one of them being that “JWST took the very first pictures of a planet outside of our own solar system.”

While we know that the JWST is the most advanced telescope to peep at the universe with and can take better photos of planets outside our solar system, saying that the JWST took the very first picture of an exoplanet isn’t factual. As stated on Nasa’s website, the first image of an exoplanet was taken back in 2004, while the JWST became operational in 2022. The first exoplanet was imaged by the VLT (Very Large Telescope), stationed in Cerro Paranal, Antofagasta, Chile.

Several astronomers and astrophysicists were quick to point out the inaccuracy of Bard, with Grant Tremblay (@astrogrant) saying, “Not to be a ~well, actually~ jerk, and I’m sure Bard will be impressive, but for the record: JWST did not take “the very first image of a planet outside our solar system.” He added, “I do love and appreciate that one of the most powerful companies on the planet is using a JWST search to advertise their LLM. Awesome! But ChatGPT etc., while spooky impressive, are often *very confidently* wrong. Will be interesting to see a future where LLMs self error check.”

In a statement given to The Verge about Bard’s inaccuracy, a Google spokesperson said, “This highlights the importance of a rigorous testing process, something that we’re kicking off this week with our Trusted Tester program. We’ll combine external feedback with our own internal testing to make sure Bard’s responses meet a high bar for quality, safety and groundedness in real-world information.”

In related news, Microsoft has partnered with OpenAI to bring ChatGPT-like functionality to the Bing search engine. Read more about it here.

Via: The Verge

Categories
Mobile Syrup

Google unveils its AI chatbot Bard and feature updates at Paris event

Following in Microsoft’s footsteps, Mountain View, California-based Google today, at its Live from Paris event, announced AI-focused updates to Search, Maps and Translate, via The Verge.

Just two days ago, on Monday, February 6th, Google announced that its ChatGPT competitor ‘Bard’ will be available to the public in the “coming weeks,” with Google CEO Sundar Pichai describing it as an “experimental conversational AI service” powered by LaMDA. The chatbot made an appearance at the event, where Prabhakar Raghavan, senior vice president at Google, said that users would be able to interact with Bard to explore complex topics, collaborate in real-time and get new and creative ideas.

Google then explained how some questions have No One Right Answer, or ‘NORA.’ This is applicable to questions like “what is the best constellation to look at when stargazing.” The answer to such questions is subjective, and hence, it has NORA. To help answer such queries, Google is introducing generative AI directly into Search results.

Soon, if you ask Google Search questions that have NORA, the new generative AI features would organize complex information and multiple viewpoints and opinions, and combine them in your search results.

Here are some of the other new announcements made across Google’s platforms, with some features releasing in the near future, and others in the coming weeks and months:

The Street View and Live View mixture, called Immersive View, is now beginning to roll out in five cities, namely London, Los Angeles, New York City, San Francisco, and Tokyo. The feature will next expand to Florence, Venice, Amsterdam, Dublin, and more.

The multi-search tool that allows users to initiate a search using an image and a few words of text is also receiving an update. The feature allows users to take photos of objects like food, supplies, clothes and more, and add the phrase “near me” in the Google app to get search results showcasing local businesses, restaurants or retailers that carry that specific item. The feature was limited to the United States, but is now rolling out globally wherever Google Lens is available. The feature will also be available on mobile web globally in the next few months.

Under Google Maps, the company is adding new features to assist EV drivers, including suggested charge stops for shorter trips, filters for “very fast” charging stations and indications of places with chargers in search results for places like hotels and grocery stores.

Further, ‘Translate with Lens’ for images is now rolling out globally. Normally, if you’d translate the text on an image, the translated text would be added on top of the image as ‘extra text,’ and wouldn’t be blended in. This would block or distort the image behind the text. Now, with a new machine learning tool called ‘Generative Adversarial Networks’ (the same tool used in Pixel phones for Magic Eraser), Google Lens can translate the text, and blend it back into the background image without distorting it, making the image retain its natural look.

ETAs and turn-by-turn navigation via Google Maps would now be visible on your lock screen. The feature is also compatible with iOS 16’s Live Activities. Google Translate will also be introducing additional context and information for certain words or phrases, starting with English, French, German, Japanese, and Spanish languages in the coming weeks. The new Google Translate design for Android will be available on iOS in the near future.

Follow the links to learn more about the new Search, Maps and Translate features.

Image credit: Google 

Source: Google, Via: The Verge

Categories
Mobile Syrup

Google’s ChatGPT competitor ‘Bard’ to go public in the coming weeks

Last week, news came out that Google is reportedly testing its own ChatGPT-like chatbot called ‘Apprentice Bard.’

Now, Google CEO Sundar Pichai has confirmed the existence of the project, describing it as an “experimental conversational AI service” powered by LaMDA.

‘Bard’ will be able to answer user inquiries and participate in conversations in a human-like manner and is being released now to “trusted testers” before it goes public “in the coming weeks.”

Bard is similar to ChatGPT in the sense that it is all-knowing. According to Google, “Bard seeks to combine the breadth of the world’s knowledge with the power, intelligence and creativity of our large language models.” The AI service gathers information from the internet to provide fresh and high-quality responses. “Bard can be an outlet for creativity, and a launchpad for curiosity, helping you to explain new discoveries from NASA’s James Webb Space Telescope to a 9-year-old, or learn more about the best strikers in football right now, and then get drills to build your skills,” wrote Google.

Unlike ChatGPT, Google says that Bard uses less computing power, and thus, can be scaled to more users. It is also using external feedback alongside its own internal testing to make sure “Bard’s responses meet a high bar for quality, safety and groundedness in real-world information.”

It is evident that OpenAI’s decision to make ChatGPT freely available to the public is what triggered Google to go all hands on deck with Bard and make it functional as soon as possible. Another industry giant, Microsoft, is reportedly integrating ChatGPT to its Bing browser, as screenshots of the leaked browser leaked last week.

According to Pichai, AI is an exciting opportunity that can help people deepen their understanding of information and get to the heart of what they’re looking for. He adds that people head to Google for quick factual answers, like “how many keys does a piano have?”

Now, however, more and more people are turning to Google for deeper insights, with questions like “is the piano or guitar easier to learn, and how much practice does each need?” This indicates that people are looking for a diverse range of opinions and not just factual answers. Pichai says that AI can be helpful in such moments, “synthesizing insights for questions where there’s no one right answer.”

He also adds that users would soon see AI-powered features in Search, something which was hinted at in last week’s CNBC leak.

Another aspect pointed out by CNBC is that Bard will have up-to-date information, unlike ChatGPT, which is stuck in the year 2021. Pichai’s blog post doesn’t allude to that. Google was also reported to be testing alternate versions of its homepage. One of the versions reportedly gets rid of the “I’m feeling lucky’ button, and replaces it with prompts for potential questions users might wanna ask. This wasn’t hinted at, either.

The company is holding an AI-focused event about Search on Wednesday, February 8th, where we’ll likely learn more about the AI developments over at Google.

Image credit: Google

Source: Google