Powered by Data - July 1, 2024

Last Week in Conversational AI

• Character.AI has introduced a new feature allowing users to engage in voice conversations with their AI characters. This move comes amid fierce competition from rivals like OpenAI, Microsoft-backed platforms, Google, and Amazon-backed Anthropic. The startup, co-founded by Noam Shazeer, creator of the “transformer” AI architecture, enables users to customize AI companions with unique personalities and values. The new calls feature has attracted over 20 million calls from 3 million users, indicating early solid engagement.
  Read More
• Meta has begun testing user-created AI chatbots on Instagram, according to an announcement by CEO Mark Zuckerberg. These chatbots, developed using Meta’s AI studio, will initially appear in messaging and be marked as AI. The early test, launched in the U.S., involves approximately 50 creators and a select group of users, with a broader rollout expected by August. It aims to assist creators and businesses in engaging more effectively with their followers and customers. This move coincides with similar advancements, such as the AI avatars introduced by Character.AI.
  Read More
• Shopify has launched its AI “Sidekick” chatbot in early access, designed to assist merchants with tasks like creating discount codes, generating reports, and suggesting blog post ideas. Initially available to merchants with English stores in North America, Sidekick aims to tailor its functionalities to better meet merchant needs. Shopify has also announced other AI-focused updates, including AI-powered product categorization, which helps merchants by automatically suggesting product taxonomy to improve listing discoverability. Additionally, an AI tool will suggest replies for customer chats via Shopify Inbox. At the same time, AI-powered image generation is now available on iOS and Android apps, with over 1 million images generated by merchants in six months.
  Read More
• Geologists have voiced concerns about potential censorship and bias in GeoGPT, a chatbot developed under the Deep-time Digital Earth (DDE) initiative, which China heavily funds. GeoGPT, aimed at aiding geoscientists, particularly in the global south, utilizes Alibaba’s Qwen model, which some claim lacks transparency and could be influenced by Chinese state narratives. Despite reassurances from DDE representatives that GeoGPT remains independent and free from censorship, critics highlight the risks associated with using a Chinese-developed platform for sensitive geoscientific data, emphasizing the need for greater transparency in the chatbot’s development and usage.
  Read More
• Quora’s AI chatbot platform, Poe, is under scrutiny for allowing users to download paywalled articles from publications like The New York Times and The Atlantic. This controversial feature, accessed via Poe’s Assistant bot, generates HTML files of entire articles from URLs entered into the interface. Legal experts argue this constitutes copyright infringement, while Quora defends it as akin to cloud storage. Publishers, including The New York Times, have expressed outrage, with some considering legal action against such practices in the AI industry.
  Read More