Google is bringing a ChatGPT-like experience to Search and it's called ‘AI Mode'
Google has announced plans to start adding 'AI Mode' to Google Search and the Chrome browser in the US from this week, as part of its 'total reimagining of search'. This new feature will provide users with a full chatbot-like experience when searching for a query, enabling them to engage in follow-up conversations and communicate in natural language, as opposed to the collection of links typically associated with Google Search.
This new development comes as Google faces rising competition in the search engine market from rivals such as Perplexity AI and OpenAI's ChatGPT, both of which offer a chatbot-like search experience.
It is also significant that Google is launching AI Mode in Search at I/O 2025, given that AI Overviews — the AI-generated brief summary of a search query — were revealed at I/O 2024.
In a blog explaining the logic behind AI Mode, Google VP, Head of Search, Elizabeth Reid wrote, 'AI Mode is our most powerful AI search, with more advanced reasoning and multimodality, and the ability to go deeper through follow-up questions and helpful links to the web'
Google says that both the AI mode and AI Overviews will be powered by a custom version of Gemini 2.5 from this week.
'AI Mode uses our query fan-out technique, breaking down your question into subtopics and issuing a multitude of queries simultaneously on your behalf. This enables Search to dive deeper into the web than a traditional search on Google, helping you discover even more of what the web has to offer and find incredible, hyper-relevant content that matches your question.' Reid added.
The company also noted that AI Mode will help users to visualise results by analysing complex datasets and creating custom-built graphs for particular queries. Google also plans to provide a more personalised experience for users in AI Mode by offering suggestions based on previous searches and providing the option to link other Google apps, starting with Gmail, to access additional personalised content.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Hindu
19 minutes ago
- The Hindu
Apple faces intense WWDC pressure; Elon Musk deletes post about Trump; Chinese hackers aim for U.S. smartphones
Apple faces intense WWDC pressure Apple is set to host its annual Worldwide Developers Conference (WWDC) tonight at Silicon Valley. While the event is largely aimed at developers, the iPhone-maker generally puts out a slew of announcements and launch updates about its operating systems, device upgrades, AI enhancements, and more. However, this year will see Apple presenting its annual event under significantly more pressure, as it faces customer anger and sharp questions over the delayed deployment of all the Apple Intelligence features it promised. The elephant in the room will be Apple's inability to release the AI-enhanced Siri assistant as previously promised, and the disappointment of Apple iPhone 16 buyers who purchased the phone in order to access this. Analysts have predicted this year's WWDC could see Apple attempting to rebuild its damaged credibility or possibly focusing on more standard updates regarding its OS. Others are expecting possible announcements about Apple integrating Google's Gemini, Perplexity, or chatbots from other companies. Elon Musk deletes post about Trump Elon Musk has deleted a post on his platform X (formerly Twitter) where he had accused U.S. President Donald Trump of being named in the Epstein Files. This referred to U.S. government files related to allies of Jeffrey Epstein, who died by suicide in 2019 while facing sex trafficking charges. Musk had posted the claim on X during a highly-public feud with Trump over the latter's controversial spending bill. The post garnered millions of views and ignited a range of reactions across the political spectrum, as Musk claimed that Trump's alleged inclusion in the list was the reason it had not been made public. However, Musk deleted the posts over the weekend. He and Trump continued to criticise each other from their respective social media platforms; Trump is an avid user and endorser of the right-wing Truth Social. The digital fight between the two billionaires also served to bring curious and amused viewers to X, with views and activity levels reportedly spiking as high-profile insults were hurled. Chinese hackers aim for U.S. smartphones While cybersecurity experts warn about digital attacks against the topmost layers of the U.S. government and military, one major vulnerability that is often left out of discussions is a user's everyday smartphone. In order to gain access to a victim's calls and communications, hackers no longer need to dupe users into clicking a link. Instead, Chinese hackers are reportedly targeting the smartphones of high-profile U.S. users in order to carry out espionage operations, according to multiple cybersecurity experts. Chinese hackers have also tried to compromise phones used by U.S. President Donald Trump and JD Vance during their 2024 campaign. Just months ago, U.S. authorities warned that a large-scale Chinese hacking campaign was trying to gain access to the texts/phone conversations of an unknown number of U.S.-based users.
India Today
32 minutes ago
- India Today
A 60-member team of mostly IIT grads plans to become the Adobe of India — with a little help from Apple
Photo and video editing aren't new tricks, even if some of the tools that make these things possible today have changed remarkably since the early days. And they seem to be constantly evolving at lightning-fast speed in the age of large language models (LLMs) and generative a time when OpenAI, Google, and virtually every other big tech company have started making these tools available to users at the flick of a button, one wonders what will happen to dedicated photo and editing apps, particularly from smaller players. Are their days numbered? Amid all the risk and uncertainty, a quietly ambitious company from Noida is charting a course to redefine editing, with its sights set on becoming the 'Adobe of the subcontinent.'advertisementAt the helm of this uphill task is Sharad Shankar, founder and CEO of AndOr Communications Pvt. Ltd., whose flagship product, LightX, is rapidly gaining traction in India and abroad. It is not only challenging the status quo but also inspiring a generation of budding developers to dream big and work hard. We sat down with Shankar to delve into the company's journey, its bold aspirations, and its deep integration with Apple's ecosystem that are helping LightX stand out in the burgeoning field of AI-driven content creation. AndOr Communications was founded in December 2015, a pivotal time in the mobile technology space. 'At that point of time, we realised that iPhones were getting really powerful in terms of processing power,' Shankar says. This foresight formed the basis for LightX. 'We realised early on that content creation is something which is going to be used widely in the coming years, and we thought of building a solution which could do the same task that Photoshop can do on a desktop.'advertisementLightX is a comprehensive suite of photo and video editing tools designed for the mobile-first creator. 'You can do all kinds of photo and video editing tasks like cutting, cropping, trimming, cleaning up, removing backgrounds, take your selfie and apply different kinds of makeup, change the expression of your face and things like that,' Shankar traditional editing, LightX provides a vast array of templates for designing graphics like cards and posters. More recently, the company has pivoted significantly towards artificial intelligence. It claims to offer unprecedented control over images, letting users change hairstyles and outfits in portrait shots, turn them into caricatures, and replace certain things and items within them, all with prompt input.'With AI, you can generate images really fast and speed up the creative process,' Shankar brains behind LightXShankar is an IIT Kanpur alumnus. 'Initially, we were providing and creating a complete mobile ecosystem for print systems. We also worked with a company based out of California to develop their complete mobile engineering system for a customer CRM based program on mobile.' One thing led to talent pool at AndOr Communications is notable. 'Most of my core 60-member team comes from IIT,' he says proudly, highlighting the strong academic foundation of his engineering team. The initial foray of LightX into the market was met with significant success. "We rolled out the first version of LightX in May 2016. It was a paid app priced differently in different territories, from USD $1 to USD $5,' Shankar recalls. 'We sold nearly 1.2 million copies at the initial launch of LightX with the global rating of 4.7 out of five and we were one of the top paid apps globally in 2016 and 2017 in the US and most of the European countries.'advertisementRecognising the evolving app economy, the company strategically shifted its monetisation model. 'A one-time purchase model was not feasible for us in the long term, so we moved to a subscription-based model where you pay USD $4 a month or USD $48 a year. It has been downloaded 7.55 million times globally since then, and we still maintain a 4.7 out of 5 rating out of 138K ratings that we have received.' This transition has ensured sustained growth and engagement in the competitive app from AppleLightX's success is intricately linked with its deep integration into the Apple ecosystem. 'Apple has been helping us a lot by featuring us during different events. It gives great visibility to our software and the things that this software can do,' Shankar visibility, Apple's technologies have been instrumental in LightX's performance. 'We have been integrating with the technologies that Apple offers on the device, especially around speeding up your workflow, because most people want to see results instantly. We have been integrating the Metal framework, and we have also been writing our own shaders on top of that.' The Metal framework, specifically, allows LightX to achieve high-performance rendering and seamless user experiences, critical for real-time image and video learning models that run locally on the device itself enhance speed while addressing privacy concerns. 'Apple has a very good ecosystem for deploying machine learning models on devices themselves. So, whatever machine learning models that we train can be easily deployed on iPhones using Core ML models, and we use their Vision frameworks for computer vision-related tasks like detecting faces and poses."LightX also leverages the App Store's promotional tools effectively. 'We have been using all the features that the Apple App Store provides, particularly the App Store events. We also have different custom product pages for different marketing channels, which helps in better discovery of the product.' This strategic use of App Store features maximises their reach and user mentions that most of the downloads they get come from organic channels through keyword search and 'it works in a way where we get a good number of users from all countries, otherwise it's a very expensive process to get downloads globally.'The future of editingShankar details several ambitious plans for LightX's AI capabilities, particularly in semantic and generative editing. 'We are working towards enhancing the experience of editing faces where, if you have a group of five people and one is looking in another direction, you can adjust and make this person look in the correct direction. Similarly, you can apply different touch-up options.'Moreover, the power of natural language prompts is set to revolutionise user interactions. 'We already have prompts on iOS which give you the power to do generative editing. We are working to build even more advanced semantic photo editing where you can just say you want to change the sky to blue, or you want to move a person backwards, correct their face, make them smile - you can do all this by typing a prompt and it will happen. It is more intuitive than finding the right settings manually."advertisementThis extends to graphic design as well. "Similarly, we are working to build designs using prompts. Suppose you have a daughter named Sarah who is 10 years old, and she loves hiking. So, you can say: create a birthday card for Sarah, she is into hiking and is ten years old, so it will automatically generate the text and theme for the birthday card. Ultimately, we are trying to make LightX a one-stop solution for all kinds of creative needs.' The idea, if you can read between the lines is, to democratise sophisticated editing tools.'We use open-source models which are based on Stable Diffusion because it takes a lot of effort in training foundation models. So right now, we are taking the foundation models as they are. We retrain the system, or you can say, distill the system for our specific needs,' Shankar he acknowledges the challenges. 'We understand that open-source models only work to a certain degree and proprietary models from say, OpenAI, are more advanced, so it becomes difficult at times to compete with those players. We are in the process of training our own models which can give us more native and India-specific content.'The ambition extends to adding multilingual support. "Eventually, we plan to integrate with Siri, and make the models understand and respond to queries in Hindi, Tamil, and Telugu to create graphics designs. We are working towards more optimisation of those channels, which can give a more native experience to the users in India. We have started the training process for different languages in which images can be generated as well. Obviously, in text there are lots of models from OpenAI and others. But we are working more towards image generation based on your command prompts."Apple's commitment to developers is a significant advantage for companies like LightX. 'Apple already provides SDKs around voice that reduce the cycle time of developing things. We use their SDKs, and they also give us prompts. We have our own LLMs at the back end which understand these prompts and then create the image. Apple helps us in terms of rolling out the features that developers can use rather than building everything from scratch. It's difficult to build each piece of the puzzle. So, they are providing us a very good intelligence system which we use to derive our own systems.' This symbiotic relationship with Apple allows LightX to focus on its core expertise in imaging while leveraging the iPhone maker's world-renowned and battling DeepfakesIn an era of increasing concerns about data privacy and the misuse of AI, Shankar outlines LightX's robust policies. 'We don't use user images for training purposes.'He elaborates on their image retention and content filtering measures. 'We delete all the images within 24 hours of acceptance and if a user wants to keep any generated photo, then only they can keep it. Otherwise, we automatically delete all kind of user input images. Other than that, we have a strict check on NSFW content. Not only do we do content filtering, once the image is generated, we analyse it further if it contains any kind of nudity. If we detect anything inappropriate, then we reject it promptly.'Compliance with global regulations is also a priority. 'Also, we do setups in GDPR. That is for Europe because they want their data not to be transferred to any other country so that there is no kind of data violation.'Competing with giantsThe rise of large language models and generative AI from tech giants like OpenAI poses a significant challenge for smaller players. Shankar accepts this reality but says LightX can still make it because of its unique advantage.'It's a concern, but we can still make it work. OpenAI will take INR 5 or INR 8 for photo editing, but players like us who are doing their own technology, building their own stack, are more specialised, and we do it in just 20 or 30 Paisa. So, we have a huge window of cost opportunity where we can innovate,' he says. This cost advantage is what gives LightX the bandwidth 'to do more innovation', which is also greatly customised and not just stand out but 'definitely win' over bigger companies.'It's challenging, but eventually, I think people who can sustain through this and can keep innovating will come out as survivors. So, we constantly focus on building our own tech stack and expect it will keep us alive for a long time.'
Mint
an hour ago
- Mint
Apple's WWDC is here. Investors are still waiting on that AI promise.
Apple's Worldwide Developers Conference kicks off Monday with muted expectations from Wall Street—and the risk of last year's unfulfilled AI promises overshadowing the conference. In 2024, the company put out its answer to 'what's Apple going to do with AI?" The announcements included a range of new features under the banner of 'Apple Intelligence," but the rollout was delayed. 'Expectations are rightfully tempered this year at WWDC relative to prior years where we had more material announcements," wrote Evercore ISI's Amit Daryanani in a note to clients last week. Like many, Apple was taken by surprise by the November 2022 launch of OpenAI's ChatGPT, the match that lit the current AI boom. Apple had long been successfully using AI throughout its products for specific features like making photos, videos and audio higher quality, but had failed to make Siri a useful conversational assistant, like the new generation of chatbots are providing. At WWDC 2024, Apple announced some new AI features that were becoming common elsewhere, such as text and image generation, and notification summaries. But the big new advance would be a redesigned Siri—a real assistant that could work across installed apps and the web, creating automated workflows from simple conversational commands. But the rollout was delayed last fall and on arrival it omitted the new Siri, for which Apple users are still waiting. Existing features have been underwhelming and were met with poor reviews. 'Apple's made this promise, this huge thing is coming that's going to change everything across their whole lineup," said prominent tech reviewer Marques Brownlee in November, as features were rolling out. 'They've been talking about it for a while, and I think that promise is starting to fade." In a March statement to Apple blogger John Gruber, an Apple spokesperson admitted that the new Siri is 'going to take us longer than we thought to deliver on these features and we anticipate rolling them out in the coming year." On its May earnings call, Apple CEO Tim Cook made similar remarks. Ben Reitzes of Melius Research said that management can't be pleased as it gears up for the conference, 'but we have no doubt they are making adjustments to the strategy behind the scenes." 'However, it is unlikely we'll hear the company outline precise aspects of its upcoming product plans until September," Reitzes said. Even if it is racing from behind, no one should count Apple out yet. Its biggest strength is its installed base of over two billion devices. Anything it puts out will instantly be in front of a huge number of people. But Apple needs to do better than it has to date to get those users engaged, and that has resulted in a leadership shake-up for this new version of Siri. A large part of the problem is self-inflicted by other Apple concerns. Privacy and security are at the heart of everything Apple does, and it would prefer that neither Apple or anyone else have access to your conversations with Siri. The best way to do this is for everything to happen on the device, not in the cloud somewhere, and this is how most of Apple's legacy AI features work. The first layer of Apple Intelligence is on-device AI, but this is hampered by the limited computing and memory capabilities of consumer devices, and results are mixed. Apple has also built its own cloud, called Private Cloud Compute, and its first order of business is that no one can see user chats. Apple wants to control the hardware and software so it can guarantee private and secure chats, but this puts it at a disadvantage to the Nvidia technology that so many competitors use. That is why, at least for now, Apple is forced to fall back on the third layer of Apple Intelligence, ChatGPT, where Apple cannot guarantee user privacy. Security and privacy may turn out to be Apple's biggest hurdles in creating the Siri that its users would like to have, and that Apple promised them a year ago. Apple wasn't the first to make a PC, smartphone, tablet or smartwatch, but it eventually made category-defining products for all of them. Making Apple Intelligence a leader will be more difficult than any of those. Write to Adam Levine at
