Microsoft Launches Multilingual AI-chat Bot Jugalbandi for Rural India

By

-

May 25, 2023

Microsoft multilingual AI-chat bot Jugalbandi rural India — Image Credits: Stock Images

Microsoft on Tuesday unveiled Jugalbandi, a multilingual chatbot powered by generative AI and accessible through the well-known messaging app WhatsApp. The bot has been created specifically to cover rural parts of India that are difficult for media to reach and don’t have access to government social programmes.

“I put this to the test when I recently visited India, and I came to two conclusions. First, the fact that we have the ability to create products that can benefit 8 billion people rather than just a select few. And rather than taking years or centuries, this dissemination will be able to happen in just a few days or weeks. With this, we also secure protection of fundamental rights,” said Satya Nadella when announcing Jugalbandi.

In partnership with IIT Madras, AI4Bharat created the chatbot. According to Microsoft, it aims to offer individualized support by comprehending user inquiries in a variety of languages, whether they are spoken or typed.

After receiving the request, the chatbot gets data from relevant programs, which are frequently only available in English, and returns it to consumers in their native tongue.

In order to provide smooth communication between users and the chatbot, Microsoft claims that Jugalbandi uses AI models from AI4Bharat and Microsoft Azure OpenAI Service. Large volumes of data can be synthesized by these generative AI technologies to produce text and other types of content.

By utilizing the databases owned by the Indian government, Jugalbandi provides consumers with a trustworthy and accountable platform while also ensuring data protection, according to Microsoft. The chatbot was introduced in April and put to the test in the village of Biwan, which is close to New Delhi, the capital of India.

Microsoft Announces Plugins for Microsoft 365 Copilot

By

Sahil Pawar

-

May 25, 2023

Microsoft Announces Plugins for Microsoft 365 Copilot — Image Credits: Microsoft

As part of its Microsoft Build 2023 announcements, Microsoft announced that its AI-powered assistant for its suite of office products, 365 Copilot, will receive plug-ins. With the update, users will be able to upgrade the assistant’s features and tailor it to meet their unique requirements.

According to a blog post by Microsoft EVP Rajesh Jha, there will be three different kinds of Microsoft 365 Copilot plug-ins, including plug-ins utilizing OpenAI’s ChatGPT technology, Teams messaging extensions, and Power Platform connectors.

There will be more than 50 plug-ins available when the early access programme debuts, including plug-ins from Atlassian and Adobe. Microsoft claims that there will be thousands of plug-ins available when they go live.

It is the organization’s most recent effort to make its AI tools more adaptable for a wider range of users and might play a significant role in how the organization sets itself apart from its rivals. Under the Duet AI branding, Google has its own AI tools geared towards productivity, although they are not yet generally accessible and do not enable plug-ins.

It is not just Copilot that is getting plug-ins; last month, Microsoft revealed that plug-ins will be coming to its Bing AI, which might be beneficial for, for instance, finding a restaurant reservation through OpenTable.

Additionally, Microsoft and OpenAI are supporting an “open standard for plug-ins that integrates across OpenAI’s ChatGPT and Microsoft copilot offerings,” Jha wrote. OpenAI recently announced support for plug-ins with ChatGPT.

Microsoft’s Bing To Start Serving As ChatGPT’s Default Search Engine

By

Sahil Pawar

-

May 24, 2023

Microsoft's Bing ChatGPT's default search engine — Image Credits: Bing

According to Microsoft Build 2023 announcements, Bing will start serving as ChatGPT‘s default search engine. Users of ChatGPT Plus will begin receiving the new search functionality today, and all free ChatGPT users will soon have access to it.

Microsoft’s VP of consumer marketing, Yusuf Mehdi, writes in a blog post, “Our outstanding cooperation with OpenAI is fundamental to our progress with the new Bing. ChatGPT now has a top-notch search engine integrated to deliver quicker and more current responses with access from the web.”

OpenAI’s AI chatbot ChatGPT will leverage search and online data that also contains citations, similar to how Bing’s chat experience is driven by GPT-4 and incorporates citations to links. Bing’s incorporation with ChatGPT is a win for Microsoft after Samsung recently elected to continue with Google despite rumors that it will switch to Bing Search.

The integration comes after years of tight collaboration with OpenAI and Microsoft’s decision to expand their cooperation with a multibillion dollar investment earlier this year.

In order to facilitate interoperability between plug-ins for Bing Chat, Microsoft’s Copilot platform, and ChatGPT, Microsoft has also committed to using the same open plug-in standard that OpenAI does for ChatGPT.

Google Releases New Generative AI Tool Product Studio for Sellers to Edit Product Images

By

Sahil Pawar

-

May 24, 2023

Google generative AI tool Product Studio — Image Credits: TechCrunch

At the Google Marketing Live event, Google unveiled Product Studio, a new tool that enables merchants to quickly create and update product images using generative AI. Product Studio will be made accessible in Google’s Merchant Centre Next, a platform that allows companies to choose how their products are shown on Google.

Product Studio aims to make it simple for merchants to produce updated product imagery without incurring additional expenses from additional photoshoots. Merchants can utilize generative AI to generate minor visual changes to product photographs with just a few prompts, such as creating a variety of fresh backgrounds for seasonal promotions.

The release of Product Studio is scheduled for the next few months for US-based vendors. According to Google, Shopify users of the Google and YouTube app will also get access to Product Studio functionalities. Merchants can utilize a variety of features in Product Studio to spice up their product picture.

With the custom product scenes function, retailers may input a product image without a background and utilize artificial intelligence to create various background settings to match different marketing campaigns or to just swap out a boring or unattractive backdrop. Merchants can also remove an existing background from an image in Product Studio to create a white background. Small or low-resolution photos can be enhanced using Product Studio to increase their quality and lessen any discernible blurring or pixelation.

Product Studio is meant for companies of all sizes, but it will likely be especially helpful to small firms and independent sellers that lack the funds to engage a professional photographer. The technology may, however, also be useful to larger businesses that need to swiftly develop fresh product pictures for advertising campaigns.

AI-Generated Image Falsely Reports Blast Near Pentagon, Tricks News Outlets

By

Sahil Pawar

-

May 24, 2023

AI-generated image falsely reports blast near Pentagon — Image Credits: CNN

On Monday, May 22, late in the evening, a number of electronic media channels in India started to broadcast stories about what appeared to be an explosion close to the Pentagon in Washington DC. While some people posted a brief news update on social media, others broadcast a picture of a thick cloud of smoke rising from a building that was probably close to the Pentagon.

Whoa, did you hear what happened this morning? An AI-generated image of an explosion at the US Pentagon spread like wildfire! The S&P 500 even took a hit, dropping 30 points in a matter of minutes. But get this – the image was fake! As in, totally bogus. And now people are… pic.twitter.com/KkYkT73xgw
— Jewels Jones (@JewelsJonesLive) May 22, 2023

Nick Waters, a digital investigator with Bellingcat, a well-known investigative news website, made an important observation not long after these images attracted widespread notice. Waters emphasized that the extensively shared photos displayed key traits typical of artwork created using artificial intelligence.

It’s also important to note that there were no other credible public accounts with images depicting the claimed explosion. The official Twitter account of Arlington Fire & EMS responded to the online rumors by issuing a message to clarify the situation and stress that there was no explosion or incident happening within or near the Pentagon.

Arlington County, Virginia is home to the Pentagon, which serves as the US Department of Defense’s main office. There is no immediate danger to the public, as the tweet expressly noted that “There is NO explosion or incident taking place at or near the Pentagon reservation.”

@PFPAOfficial and the ACFD are aware of a social media report circulating online about an explosion near the Pentagon. There is NO explosion or incident taking place at or near the Pentagon reservation, and there is no immediate danger or hazards to the public. pic.twitter.com/uznY0s7deL
— Arlington Fire & EMS (@ArlingtonVaFD) May 22, 2023

The images were broadcast on television by Republic News at around 7:45 pm, and a “strategic expert” was summoned to explain the incident. Messages like “RT reports explosion near Pentagon,” “Images show huge plumes of smoke,” and “More details awaited on Pentagon situation” were featured at the bottom of the screen.

Times Now Navbharat, News18 Madhya Pradesh, and First India News all shared a newsflash with their followers on social media to let them know about the occurrence. Zed News initially made a video report about the incident, however it was later taken down.

Microsoft Announces AI Personal Assistant Windows Copilot for Windows 11

By

Sahil Pawar

-

May 24, 2023

Microsoft AI personal assistant Windows Copilot — Image Credits: GeekWire

Microsoft declared on Tuesday that its artificial intelligence-powered chatbot Windows Copilot would be included to the Windows 11 home screen. Similar to how Cortana, Microsoft’s existing AI-powered virtual assistant, is presented, Windows Copilot will act as a central, text-based AI assistant that can be accessed from the Windows taskbar.

Users will be able to launch programmes like Adobe, Spotify, Teams, computer settings, and more while using Copilot’s chat function in an on-screen sidebar, thanks to the integration with a range of other apps. “The button is prominently displayed on your taskbar and is easy to identify and use,” according to Microsoft’s Panos Panay, chief product officer for Windows and devices.

“The Windows Copilot sidebar, which serves as your personal assistant when open, is consistent across all of your open windows, apps, and programmes. Every user becomes a power user as a result, making it easier for you to take action, personalize your settings, and connect to all of your favorite apps at once,” they added.

To assist users in completing activities, Windows Copilot will be able to explain, rewrite, or summarize copied text or documents. Users will be able to enter up to 2,000 characters in the chat field, according to a video on Microsoft’s website.

The Bing search engine and ChatGPT will both have plugins for Copilot that will further the integration of AI-powered technologies into Windows programmes. Additionally, it will enable programmers to include apps inside Windows, thereby improving user experience and increasing interest in Windows programmes.

Beginning in June, Microsoft intends to make Windows Copilot accessible to Windows 11 customers so they can preview the upcoming virtual assistant. Recently, the tech corporation previewed an updated AI “copilot” for Microsoft 365, a collection of Word documents, PowerPoint presentations, Excel spreadsheets, and Outlook emails.

Google Introduces AI-driven Feature Assisted Vision in its Lookout App

By

Sahil Pawar

-

May 24, 2023

Google AI-driven feature Assisted Vision Lookout app — Image Credits: CNN

Google has introduced a new AI-driven feature in its Lookout app called Assisted Vision. Visually impaired people can use this AI feature to ask several questions about the details of an image.

Google unveiled the “Image Q&A” function in its accessibility application Lookout and gave a brief demonstration of it during some major announcements at Google I/O 2023 recently.

Visually impaired can ask any question to understand what is in the image and the artificial intelligence provides them with a clear answer every time. The purpose of this feature, and of Lookout in general, is to help visually impaired people understand a photo or drawing.

An image needs to be uploaded to the app, before asking the AI to describe it. The latter will execute and explain what can be seen on the screen. This first description is effective and might suffice if one just wants a quick summary. However, those who need to go into more detail can just ask a question, in writing or out loud.

Describing an image is no longer a very complex challenge for a well-trained artificial intelligence. However, Google has successfully taken the concept one step further to make it more practical and beneficial for people with disabilities.

Infosys Unveils Infosys Topaz, an AI-first Set of Services, Solutions, and Platforms

By

Sahil Pawar

-

May 24, 2023

Infosys unveils AI-first Infosys Topaz — Image Credits: Mint

On Tuesday, Infosys, a leading provider of information technology, unveiled Infosys Topaz, a new product that blends data analytics, artificial intelligence, and generative AI. Infosys Topaz is an AI-first set of services, solutions and platforms using generative AI technologies.

Infosys claimed to have created an AI-first core with the potential for over 12,000 use cases using its own applied artificial intelligence framework. Infosys Topaz democratizes data and information to benefit more ecosystem players, empowering them to develop business models, AI-driven goods and services, and new sources of income.

The goal of Infosys’ AI service is to increase productivity throughout the company. For business agility, it makes use of intelligent tools, platforms, and autonomous software engineering.

“By reinventing user profiles, data architecture, and engineering plans for the future, Infosys Topaz drives organization-wide synergies. In addition, by using enterprise, it aids in the development of self-supervision skills,” said the company’s statement.

“Infosys Topaz is assisting us in maximizing both our own and our clients’ human potential. While businesses are eager to secure their future growth, we are witnessing significant interest from our clients in programmes that increase productivity and efficiency. The strength of generative AI platforms and data solutions brought by Infosys Topaz has greatly helped our own business processes,” said Salil Parekh, CEO & MD of Infosys.

Koo Introduces BERT-based Pretrained Model KooBERT

By

Sahil Pawar

-

May 23, 2023

Koo BERT based pretrained model KooBERT — Image Credits: Koo

Koo has introduced KooBERT, a masked language model trained on data from the multilingual micro-blogging social media platform Koo India. This BERT based pretrained model was built in collaboration with Koo India and AI4Bharat.

In his LinkedIn post, Head of Machine Learning & AI Harsh Singhal said, “KooBERT is a testament to our commitment to inclusivity, diversity, and multilingualism in Al. Trained on a large corpus of Koos 10+ Indian languages, it’s a significant leap forward in democratizing Al for millions of non-English speakers in India and around the world.”

On the Koo platform, there are microblogs (Koos) which are limited to 400 characters and available in multiple languages, including assamese, Bengali, English, Gujarati, Marathi, Oriya, Punjabi, Tamil, Hindi, Kannada, Malayalam, and Telugu. The model was trained on a dataset that contains multilingual koos from Jan 2020 to Nov 2022 on masked language modeling tasks.

This model can be used to perform downstream tasks like toxicity detection, content classification, and more for supported Indic languages. It can also be used with sentence-transformers library for the creation of multilingual vector embeddings for other uses.

As with any machine learning model, KooBERT does have limitations and biases. This model has been trained on Koo Social Media data and may not generalize well for other domains. It is also possible that the model may have biases in the data it was trained on, which can affect its predictions.

Naval Launches Generative AI-based App Airchat that Resembles TikTok and Clubhouse

By

Sahil Pawar

-

May 23, 2023

Naval Generative AI-based App Airchat Resemble TikTok — Image Credits: Airchat

New AI-based application called Airchat is a mix of different social media apps, including TikTok and Clubhouse. The presence of generative AI to improve the user experience is what makes Airchat an interesting addition to social apps.

Social media isn’t social – it’s performative.

Where’s the chitchat, the banter, the easy laughter?

We made something new.

Push-to-talk, whenever you want.

With perfect transcripts and AI art.

A dinner party in your pocket.

Follow @getairchat

Never feel alone again. pic.twitter.com/V9urTOVf4u
— Naval (@naval) May 21, 2023

Airchat gives you the option to “push to talk” whenever you want. The software allows users to interact with one another through Clubhouse podcasts or videos. For many topics, the app features monitored discussion rooms. People can enter the chat room of their choice and talk to the speakers inside. Alternatively, they can just listen in on discussions.

Each chat in an Airchat room is broken up into several messages or sections, which is what makes them unique. In a lengthy discussion, users have the choice of replying, liking, or listening to a particular message. There is more than just audio and video on Airchat. There are also text-based rooms that use AI to suggest content to the users of the app.

The software also places a strong emphasis on user privacy. There are open, public, and private rooms. The privacy setting for a room can be chosen by the room’s developer.

Generative AI is heavily emphasized in Airchat. An AI-generated transcript is included with each conversation on Airchat, whether it be audio or video. Transcript translations are now available on the app. There is still more AI engagement. For their discussions, users can create custom image backdrops. The picture will coincide with the chat’s transcript.

On iOS, Airchat is presently accessible as a closed beta. Those who are interested can sign up for the app’s waitlist here. Currently, the app is being developed on Android.