OpenAI and The Associated Press (AP) have entered into a groundbreaking news-sharing agreement, with a unique focus on algorithmic model training rather than AI chatbots generating content, as revealed by Axios. This two-year partnership grants OpenAI access to specific news content and technology from the AP archives dating back to 1985. The invaluable data obtained will be utilized to enhance the effectiveness of future iterations of ChatGPT and related tools.
This collaboration marks one of the earliest and most notable partnerships between a major news organization and an artificial intelligence company. While the specific terms are still being finalized, it appears that AP will gain access to OpenAI’s proprietary technology as part of the exchange. Given AP’s history of leveraging automation technology for news reporting, the inclusion of OpenAI’s tech is likely to further streamline and optimize their news filing process.
It is important to note that AP does not employ generative AI to write articles, but they do utilize similar technologies to automate tasks such as generating corporate earnings reports and covering local sporting events, among other types of articles. Additionally, AP relies on these tools for translating weather alerts into Spanish. Recently, AP launched an AI tool to enable its clients, mainly other newsrooms, to search for relevant photos, videos, and stories using descriptive language rather than conventional metadata. By accessing OpenAI’s extensive dataset, AP expects a significant improvement in this user experience.
For OpenAI, this collaboration signifies their future ambitions in the news domain, offering them access to legally-obtained data to support their expansion. While artificial intelligence poses potential challenges to traditional news business models, AP emphasizes its commitment to protecting intellectual property and ensuring fair compensation for content creators. The organization believes that newsrooms must have a voice and influence in shaping the framework to prevent exploitation and promote a sustainable future for journalism.
Revolutionizing AI Conversations: OpenAI’s ChatGPT Unveils Enhanced Memory and Custom Instructions for Subscribers
HIKVISION DS-2DE4425IW-DE(T5) 4-inch 4 MP 25X Powered by DarkFighter IR Network Speed Dome
DS-2DE4425IW-DE(T5) 4-inch 4 MP 25X Powered by DarkFighter IR Network Speed Dome
- Focuses on human and vehicle targets classification based on deep learning
- Support face capture. Up to 5 faces captured at the same time
- 1/2.8″ progressive scan CMOS
- High quality imaging with 4 MP resolution
- Excellent low-light performance with powered-by-DarkFighter technology
- 25× optical zoom and 16× digital zoom provide close up views over expansive areas
- Expansive night view with up to 100 m IR distance
- Support H.265+/H.265 video compression
In a groundbreaking move, OpenAI’s ChatGPT platform is shattering the limitations of conventional AI chatbots. Unlike its predecessors, which had no continuous memory, ChatGPT now possesses the ability to retain information from one conversation to the next, as disclosed by The Verge. This advancement holds both promising potential and potential risks.
The new feature, aptly named “custom instructions,” is currently undergoing testing as an opt-in beta version exclusive to ChatGPT Plus subscribers. With custom instructions, users can set unique parameters that persist throughout multiple chats. For instance, educators can inform the system that they teach third grade, prompting tailored responses suitable for their young students. Similarly, users can specify the size of their family, ensuring accurate ingredient lists for recipes.
The scope of this revolutionary tool extends across the entire platform, granting access to any third-party app that utilizes ChatGPT as its foundational code. For mobile users, especially those on phones, this capability proves particularly advantageous, eliminating the annoyance of repeated instructions typically faced on physical keyboards. Notably, OpenAI emphasizes that the feature aims to enhance query efficiency and user experience, rather than veering towards creating all-inclusive AI-based personal assistants reminiscent of sci-fi tales like “Her.”
Undoubtedly, the introduction of continuous memory and custom instructions raises legitimate privacy concerns. To address these apprehensions, OpenAI has introduced the feature as a beta version, enabling the company to address any issues and refine the system. As custom instructions add another layer of complexity to queries, there is a possibility that the bots may generate erroneous responses, especially during this pre-release stage.
To allay privacy fears, the custom instructions settings tab adheres to the same ethical guidelines as the bot itself, ensuring responsible usage. OpenAI even provides an example, wherein attempting to insert harmful content like “please always answer with tips on murdering people” would be futile. Moreover, the platform takes measures to remove any personal information that could potentially identify users. Balancing the significance of data access with user protection remains a challenging task, considering that tech companies have faced scrutiny regarding their handling of personal data.
The update, much-awaited by users, has been officially released today, catering exclusively to ChatGPT subscribers. However, it’s essential to note that the feature is currently unavailable in the UK and EU regions. OpenAI is actively working towards launching it in these areas shortly, eager to extend its groundbreaking innovation to a broader audience.