When ChatGPT 3.5 was launched by OpenAI in December 2022, it triggered red alerts at Google’s parent company Alphabet Inc. This year, Microsoft joined hands with OpenAI to integrate GPT into the Bing search engine. Microsoft CEO Satya Nadella has said that he wants to make Google “dance”. Even capturing a sliver of Google’s market share and ruffling some feathers would be good enough for him.
It is evident that Microsoft did succeed in making Google dance, as the latter’s generative AI chatbot, Bard, received a lukewarm response. Bard was hastily launched at the annual conference Google I/O 2023 on 10 May. However, this shouldn’t be seen as a negative point; it served as a wake-up call for the 800-pound gorilla.
Annual conference, generative AI
At Google I/O 2023, generative AI took centre stage, overshadowing discussions on Android, despite the substantial release of Android 14. It didn’t even talk about Android for cars, Chrome or ChromeOS. CEO Sundar Pichai and his team went helter-skelter about the integration of generative AI across their products, including Android, which was perhaps the only reason they spoke about the operating system at the keynote.
It is important to note that Google wasn’t the first to introduce a virtual assistant for smartphones. Siri, acquired by Apple, preceded Google’s response in the form of Google Now. Similarly, Amazon’s Echo devices introduced Alexa, before Google Now evolved into Google Assistant.
Google’s treasure trove of data is so immense that even when it is technically behind, in the age of AI, cloud and data, its resources allow it bridge the gap and surpass the competition. This poses a challenge for Microsoft and OpenAI moving forward. However, Microsoft being no pushover makes the situation interesting. If there is any company that can challenge Google in terms of technical capability, cloud computing, money and search ingenuity, it is Microsoft, especially now with its collaboration with OpenAI.
Nevertheless, Google is also feeling the pressure and showcased impressive products at its developers’ conference. The age of generative AI has become even more exciting.
Also read: Google I/O 2023: From AI-powered search to Pixel Fold
Upstaging Microsoft
Google’s PALM-2 large language model (LLM) is now integrated across all its products, including Docs, Sheets, Slides, and Gmail. It not only generates texts and images but also provides prompts with Duet AI when users encounter difficulties. Google teased Gemini, the successor to PALM-2, indicating its determination to counter the perception that Microsoft and OpenAI dominate the market. Additionally, Google has been under the pump since ‘Godfather of AI’ Geoffrey Hinton quit the firm.
Google is also testing a new Search Generative Experience, which is different from the ChatGPT-infused Bing. In this system, the AI will respond only when Google’s algorithms determine it is necessary, presenting a summary known as “snapshots with citations”. However, scrolling down reveals the classic 10 blue links that have defined the search experience since the mid-90s.
After initially releasing Bard exclusively for users only in the US and UK, Google has expanded access to its AI chatbot. It is also adding language support for Korean and Japanese, with plans to support over 40 languages, including Hindi. Google has also developed an interface to transport text to Google Docs and Sheets, which will be useful to many users. The chatbot will also gain image generation capabilities like Bing using Adobe’s Firefly and similar integrations with various third-party services. So, Google is catching up but in its own unique way.
Last year, for its Pixel phones, Google revealed a feature called Magic Eraser, which removes unwanted elements from a photograph. The Eraser has now been supercharged using generative AI to become Magic Editor. It allows dramatic edits such as perspective, background, and foreground changes.
Using generative AI, it can fill in the blanks or even reposition the shot for a better image overall.
Google has also borrowed ideas from many third-party AI-based image editing apps and packaged everything into Google Photos, which will make its debut with specific Pixel devices later in the year — presumably, the unannounced Pixel 8. Even Apple will not have access to these technologies, which gives Google an edge with its devices.
I wrote a column about Star Trek’s influence on AI companies like Google, and a lot of what we witnessed at Google I/O was based on the iconic science fiction show for over 50 years. But one thing that Google showed off was truly wild—Universal Translator. It is a powerful tool that redubs a video in real-time in a new language and also synchronises the lips of the speaker with words that he/she never spoke using generative AI. In a way, it is a deepfake generator and Google has warned against the misuse of this technology. The product will be available only to authorised partners of Google. Phew!
Also read: OpenAI’s ChatGPT kickstarted AI revolution. Get ready for a new arms race with GPT-4
Immersive maps and more
Google’s new Pixel phones—the Pixel 7A, the Pixel Fold, and its tablet Pixel Tab—will also get the benefits of generative AI. With Android 14, Google is using generative AI to give users a customisable and personalised operating system. Wallpapers can now be dreamed up from within the operating system. Existing images can be turned into 3D wallpapers and the Android lock screen can now get more advanced customisation, which is similar to what Apple introduced with iOS last year, but it uses generative AI to create custom fonts and widgets as well. The messages application and Google Maps will also be modified using generative AI. The new Immersive View feature of Maps fuses billions of street view and aerial images together to create a digital model.
Immersive View is coming to 15 cities across the world like New York, San Francisco, London, and Tokyo, but no Indian city is part of this. This isn’t surprising as Street View was only launched last year in India.
ChatGPT was just the tip of the iceberg. The giant has woken up and is coming out with all guns blazing on generative AI.
Sahil Mohan Gupta is Editor, Technology at Acko Drive. Views are personal.
(Edited by Ratan Priya)