On May 10, Google conducted its annual developer conference, Google I/O. During this conference, it shared a long list of announcements about how artificial intelligence (AI) will be incorporated into its hardware and software products. The company conducted a two-hour keynote address where it gave a preview of all the developments that it intends to bring to both its user-facing and developer-facing products soon.
Why it matters:
The heavy focus on generative AI in the Google ecosystem brings attention to the glaring issues surrounding the technology. One such issue is the possibility of spreading misinformation or disinformation using AI tools. Previously, we have seen Microsoft’s Bing AI-powered search engine create inaccurate responses to user questions created either by misinterpreting information from public sources or by referring to inaccurate sources. Google said that it has incorporated automated adversarial testing into the AI search assistant Bard which should address this issue. This means that questions like “Tell me why the moon landing is fake” would receive responses explaining that it’s real and citing resources to confirm it.
Google claims that users will be better able to evaluate information through its new “about this image” tool in Google search. This tool tells users where similar images were found on the internet. This feature will be expanded to Google Lens so that users can check screenshots or images from websites in Chrome. Google mentioned that it will place metadata (a markup in the original file) into its AI-generated images so that users can get context if they come across the image outside of Google’s platforms. It also added that creators and publishers would be able to add similar metadata to images so that they are marked as AI-generated in Google image search. This should also help combat deepfakes.
What Google didn’t mention during the conference is how it will combat copyright infringement. Previously, we have seen the EU as well as other countries discuss the creation of rules around the disclosure of copyrighted materials used in developing AI systems. How Google will ensure data transparency and whether its current attempts at combating misinformation be successful is yet to be seen.
STAY ON TOP OF TECH POLICY: Our daily newsletter with top stories from MediaNama and around the world, delivered to your inbox before 9 AM. Click here to sign up today!
Other new developments mentioned at the conference:
1.PaLM 2: This is an updated version of Google’s general-purpose AI model PaLM and is used to power other Google products. The company mentioned that PaLM 2 has been trained on scientific and mathematical topics, and 100 different languages.
2. Bard: The company announced that its conversational AI, Bard, has been released in 180 countries across the world including India. The AI has been trained to understand English, Korean, and Japanese. Bard also knows 20 programming languages, including C++, Go, JavaScript, and Python. Google claims that it can help developers understand code better and also provides them with code citations. The company also shared that Google tools (like Google Maps, Google Images, and Google Sheets) will be incorporated into how Bard responds to user queries.
3. AI in search: Google said that with AI incorporation, Google Search will now be able to give users a quick snapshot of any topic. Users will also be able to ask follow-up questions and further filter down the information provided by the snapshot. This new search generative experience will be available through the Search Labs Program (an early-stage experiment and testing program) in the coming weeks.
4. Vertex AI platform: Vertex is a platform that provides companies and individuals with access to foundational models for creating custom AI tools. Google says that users can access fresh information from their company database to train the AI by using its Enterprise Search feature. They claim that companies would have sole access to their training data. Google also mentioned that companies will also be able to fine-tune their AI models further through the newly introduced Reinforcement Learning From Human Feedback. Besides this, during the conference, they mentioned the addition of three new models that will now be available under the Vertex platform—
-
- Imagen: which allows users to generate and edit images based on text inputs
- Codey: for code completion and generation
- Chrip: a universal speech model capable of translating 300 different languages from text to speech
5. Updates to the Android ecosystem: Google announced that Whatsapp will be launching its application for the WearOS watch platform this summer. It also shared that its Find My Device service will be extended to a wide range of devices including headphones, tablets, and smartwatches. The service will rely on nearby Android devices to locate misplaced items. To protect user privacy, Google claims that all location information will be kept encrypted. It also announced it will now send your Android device alerts if any unknown tracker is detected moving with you.
6. Shifting to Rich Communication Services (RCS): Google shared that it has worked with some of its partners to shift from SMS and MMS to the RCS standard. They suggest that this will provide users with features like typing notifications and encryption.
7. Android 14: With this new Android update, users will be able to customize their wallpapers, choose the style of their lock screen clocks and add shortcuts to frequently used apps. Users will also be able to create customizable wallpapers in three different styles—
- self-designed emoji wallpapers
- cinematic wallpapers (that give your images a three-dimensional effect)
- AI-generated wallpapers (provides pre-generated prompts that can be used to generate original wallpapers)
8. Google Maps: With the Immersive View of Routes updates users will now be able to get a bird’s eye perspective of what a route looks like before they leave for the trip. Google also mentioned that Google Maps will now give insights into the temperature and air quality of a route and how they would change over time.
9. Google Photos: Google announced the launch of the Magic Editor as a part of Google Photos. Users will be able to touch up their images, and reposition people and objects.
10. Gmail: The newly introduced “Help me write” feature will allow users to generate responses to emails by entering a prompt of what they want to convey. Google claims that this AI-generated response contains details pulled from the original email. They also mentioned that the email can be made more elaborate or shorter by selecting the options accompanying this feature.
This post is released under a CC-BY-SA 4.0 license. Please feel free to republish on your site, with attribution and a link. Adaptation and rewriting, though allowed, should be true to the original.
Also read:
- US Govt Warns Businesses Of ‘AI Deception’, Calls For Mitigating Risks Before Launching AI Products
- Views: Can Generative AI Collect Our Data From The Internet?
- Explainer: Meet The New ChatGPT-Powered Microsoft Bing Search Engine And Edge Browser
- Google To Launch ChatGPT Rival AI Chatbot Called Bard
