Recent Advances in Technology
Summary of Recent Advances in Technology
Summary of Recent ChatGPT
Updates Support for Image and Voice
Open AI has introduced new features in ChatGPT, enabling users to interact with it using both voice and images. This provides a more intuitive interface for users to have voice conversations or show ChatGPT what they are talking about. Voice and images give you more ways to use ChatGPT in your life.
Take pictures of landmarks while traveling and discuss interesting aspects in real time when you get back home. Snap photos of your fridge and pantry to figure out what’s for dinner and ask follow-up questions for step-by-step recipes. After dinner, help your kids solve math problems by taking pictures, circling the questions, and sharing hints with both of you.
Open AI notes that voice and image capabilities will be rolled out to Plus and Enterprise users in the next two weeks. Voice will be available on iOS and Android, and images will be available on all platforms. Users can now show ChatGPT one or multiple images, enabling tasks like troubleshooting a barbecue that won’t start, planning meals by exploring items in the fridge, or analyzing complex charts for work-related data.
Customizable Interactions with ChatGPT
Open AI has introduced a new feature allowing users to customize interactions with ChatGPT based on their preferences. While currently available for Plus users, Open AI promises to make it available for all users in the near future. This new suggested replies feature in ChatGPT enhances the natural flow of conversations. It functions like an AI assistant, understanding the context of the conversation and providing appropriate responses. For Plus users, integration with GPT-4 takes the chat experience to a new level, ensuring that every interaction is powered by the latest AI technology. Conversations are more accurate, coherent, and context-aware. Additionally, for Plus users, ChatGPT now offers enhanced analytical capabilities. This upgrade allows users to upload multiple files simultaneously, enabling analysis of data from various sources, whether it’s handling code, papers, or datasets. ChatGPT serves as a versatile, intelligent analysis partner.
Co
Open AI has addressed the issue of frequent logouts. ChatGPT will now consistently provide uninterrupted assistance for tasks like brainstorming, information retrieval, and creative conversations. To enhance efficiency, keyboard shortcuts have been introduced. These shortcuts allow users to perform tasks more efficiently. For example, pressing Control + Shift + C can copy the most recent code block, and pressing Control + Shift + L grants access to a complete list of shortcuts.
Revolutionary Updates
These revolutionary updates signify the future of conversational AI. ChatGPT has evolved into a flexible and indispensable tool, fundamentally changing how we interact with technology. From personalized commands to seamless integration and enhanced analytical capabilities, ChatGPT stands at the forefront of AI development.
Voice Input Functionality
Similar to voice assistance on phones, ChatGPT now allows users to press a button, ask a question, convert it to text, generate an answer, convert it back to voice, and play it for the user. This interaction method is considered more natural and convenient. And thanks to Open AI’s technical advantages, the quality of answers is higher. Open AI has also developed a new text-to-speech model that can generate voices similar to a few seconds of sample speech. Users can choose from five options for ChatGPT’s voice.
Image Input Functionality
Similar to Google Lens, users can take pictures of things they are interested in and upload them to ChatGPT. ChatGPT will attempt to identify what the user wants to ask about and provide a corresponding response. Users can also use drawing tools in the app to express their questions or communicate in conjunction with voice or text input.
Multi-turn Conversations
One of ChatGPT’s strengths is its ability to engage in multi-turn conversations rather than one-off searches. If users are unsatisfied with an answer or want more information, they can continue to query ChatGPT for more accurate and comprehensive answers.
Limitations and Privacy Concerns
However, image search does have some potential issues, such as when dealing with images of individuals. Open AI states that they limit ChatGPT’s ability to analyze and directly evaluate individuals to ensure accuracy and protect privacy.
Open AI acknowledges that the expanded functionalities in ChatGPT still have some limitations. They admit that the model may still have potential visual confusions, misidentifying certain content, and may not perfectly recognize non-English languages. The company states that they have conducted risk assessments on the new features from extreme scenarios and a purely scientific validation perspective. The current stance is still to advise cautious use, especially in high-risk or highly specialized contexts like scientific research.
Bing Search Integration
ChatGPT can now perform web searches using Microsoft’s Bing search engine. Not limited to data before September 2021, ChatGPT Plus and Enterprise users can use the browsing feature, which will soon be expanded to all users. To enable it, select ‘Browse with Bing’ in the selector under ChatGPT.
Caution Regarding ChatGPT Results
Open AI acknowledges that the expanded functionalities in ChatGPT still have some limitations. They admit that the model may still have potential visual confusions, misidentifying certain content, and may not perfectly recognize non-English languages. The company states that they have conducted risk assessments on the new features from extreme scenarios and a purely scientific validation perspective. The current stance is still to advise cautious use, especially in high-risk or highly specialized contexts like scientific research.
Conclusion
Given the privacy issues encountered in the development of the Bis app, Open AI states that they have taken technical measures to limit ChatGPT’s ability to analyze and directly describe human objects. Because results from ChatGPT are not necessarily accurate, AI systems should respect individual privacy. While there are still various shortcomings, Open AI emphasizes in its marketing materials that ChatGPT can now see, hear, and speak. Of course, not everyone may agree with this anthropomorphic-leaning hype. Dr. Sasha Luchon, an AI researcher at Hugging Face, tweeted, ‘Stop thinking of AI models like humans. ChatGPT can’t see, can’t hear, and can’t speak. It can only be integrated with various sensors to receive and send information in a way different from humans.’ While ChatGPT and its underlying AI models are far from being human, if the results announced this time are true, it at least represents a significant enhancement in the capabilities of Open AI’s virtual assistant. Additionally, Open AI emphasizes the reason for the delayed release: ‘We believe in gradually rolling out our tools so that we can continuously improve and refine risk mitigation measures over time and also prepare everyone for more powerful AI systems in the future.’
