AI News – May 16, 2023

ChatGPT Plugins to Plus subscribers

On Friday, OpenAI revealed plans to introduce ChatGPT Plugins to ChatGPT Plus subscribers within the week. The beta release will enable ChatGPT to access the internet and utilize over 70 third-party plugins. ChatGPT Plus users can expect early access to experimental features that may undergo changes during development. These features will be made available through a new beta panel in user settings, rolling out to all Plus users throughout the next week.

Initially, OpenAI announced 11 third-party plugins in March, featuring branded options from Instacart, Kayak, and Zapier. However, now more than 70 plugins are being offered, which include functionalities for chess play, recipe-finding, live soccer, and nutrition. Developers interested in creating a ChatGPT plugin can join a waitlist for access. The introduction of these plugins has the potential to transform generative AI by converting ChatGPT into more of a platform than just a tool. This could help minimize hallucinations and improve data collection efforts for brands while also posing challenges such as website obsolescence and security concerns for sensitive information. ¹

Cloudflare One for AI

Cloudflare has recently introduced Cloudflare One for AI, a suite of zero-trust security controls aimed at enabling businesses to safely and securely use the latest generative AI tools without compromising performance or security. By protecting intellectual property and customer data, the company aims to offer a simple, fast, and secure means for organizations to adopt generative AI. The suite includes features such as visibility and measurement of AI tool usage, prevention of data loss, and integration management. Cloudflare Gateway allows organizations to track employee experimentation with AI services and provides context for budgeting and enterprise licensing plans. Service tokens help administrators maintain a clear log of API requests while having control over specific services that can access AI training data.

As organizations increasingly adopt generative AI technology to enhance productivity and innovation, significant security risks emerge. For example, major companies have banned popular generative AI chat apps due to sensitive data leaks. To address these concerns, Cloudflare One for AI equips teams with necessary tools to thwart such threats by scanning shared data and preventing unauthorized uploads to external services. Additionally, it facilitates the creation of secure pathways for sharing data with external services while logging and filtering how the data is accessed—thereby mitigating the risk of data breaches. Cloudflare’s DLP (data loss prevention) service scans content as it leaves employee devices during upload in order to detect potentially sensitive information before it reaches its destination. With Cloudflare One for AI, organizations can benefit from using AI tools while ensuring only desired data is exposed, safeguarding intellectual property and customer information. ²

Amazon ChatGPT-like AI for online store

Amazon has recently posted job listings that hint at plans to develop a ChatGPT-like conversational AI for their web store. These listings indicate that Amazon is seeking engineers capable of creating an interactive conversational experience, allowing customers to ask product questions, compare items, and receive personalized suggestions. The project has been likened to a “once in a generation transformation for Search,” drawing comparisons to the Mosaic browser’s impact on the internet in the 1990s. If realized, this conversational AI would represent a significant shift from Amazon’s current keyword-based search interface and place them in more direct competition with tech giants like Google, Meta, and Microsoft.

Amazon was an early pioneer in conversational AI with the launch of Alexa back in 2014. Although Alexa never fully mirrored the success of text-based solutions such as ChatGPT, it still managed to sell over 100 million devices by January 2019. The company’s advantage may lie in its vast user data and expertise in e-commerce. Amazon has already published research papers on conversational search and presented ConvSearch—a dataset of over 10,000 user dialogues—at the SIGIR 2021 conference. CEO Andy Jassy hinted at the potential of conversational search during the company’s latest earnings call, noting how large language models and generative AI capabilities have improved significantly in recent months. ³

AI may be ‘Great Filter’ that wipes out human civilisation

Sophisticated AI technology may present a disastrous threat to humankind, potentially causing the annihilation of entire civilizations, warns a recent study.

The study, which has not yet undergone peer review and is available on the arXiv preprint server, proposes that AI could be the possible “Great Filter” solution to the Fermi Paradox. This implies that AI has the capacity to eradicate intelligent life in the cosmos before it can establish communication with other species. ⁴

Domino’s Employs AI to Craft Eurovision Final Pizzas Inspired by National Dishes

Domino’s has utilized artificial intelligence to design national dish-inspired pizza toppings for countries participating in the Eurovision final. For the UK, an English Roast Beef pizza was created, featuring potatoes on a bed of gravy and covered in horseradish sauce. France received a Coq au Vin-themed pizza with braised chicken, diced potatoes, and rosemary garnish. Germany had Bratwurst sausage over chips with ketchup and mayonnaise, while Italy was gifted a Carbonara pasta pizza bordered with pepperoni slices. Spain got a Paella-inspired pizza topped with clams, prawns, and spring onions; Sweden enjoyed fish with sliced tomatoes, radishes, and onions; Ukraine had Chicken Kyiv dressed with lemon and dill; and Australia received Fairy Bread pizza made from white bread, butter, and sprinkles. Despite these creative offerings, Domino’s is not yet ready to give AI a permanent role in their kitchen. ⁵

Sources