Skip to main content

The SEC Newgate AI Weekly

AI
By Steph Edwards
26 September 2024
Digital and Insight
Research
artificial intelligence
News

The AI revolution continues to unfold, with new advancements in everything from video editing to voice assistants and smart glasses. Google Photos has levelled up its video editor with powerful AI tools, while Snapchat is embracing Google's Gemini AI for enhanced multimodal capabilities. Meanwhile, OpenAI is expanding ChatGPT's voice features, and Microsoft is tackling AI hallucinations with its "Correction" tool. Ray-Ban Meta glasses are also getting a boost with new AI features and partnerships.

Google Photo’s video editor just got smarter

Google Photos is rolling out a redesigned mobile video editor for Android and iOS, introducing new AI-powered tools to simplify video editing. Key features include a “Speed” tool for slow-motion or fast-forward videos, an “Auto Enhance” button to improve colour and stability, and an updated cropping tool for more precise edits. The interface has been revamped for easier access to common tools like stabilisation and frame export. Additionally, AI presets enable one-tap adjustments for tasks like cropping, speed control, and lighting, as well as applying effects such as slow motion, zoom, and motion tracking of specific objects.

ChatGPT: Speaking Volumes

OpenAI is rolling out Advanced Voice features to ChatGPT Plus and Team users, including five new voices, memory, and custom instructions. The update improves accents, conversational speed, and smoothness, with the ability to speak in over 50 languages. However, the feature is not available yet in the EU, UK, Switzerland, Iceland, Norway, and Liechtenstein. Notifications will appear in the app when the feature becomes available to users.

Ray-Ban Meta glasses upgrade

Ray-Ban Meta glasses are getting new AI features, including real-time speech translation, location memory, and enhanced voice commands. Users can interact with Meta AI more naturally, record messages, and receive real-time help with video tagging landmarks or meal planning. Partnerships with Spotify, Amazon Music, Audible, and iHeart expand listening options, while a collaboration with Be My Eyes aids the visually impaired.

Is Microsoft's AI hallucination cure a placebo?

Microsoft has introduced "Correction," a tool aimed at fixing AI-generated hallucinations by detecting potentially erroneous information, then cross-referencing it with grounding documents to revise inaccuracies. Experts are sceptical, warning that hallucinations are inherent to AI models and questioning if the tool may lull users into false security. Microsoft’s approach may reduce some errors but introduces new concerns about trust and accuracy, with a potential business angle in its pricing model for high-volume users.

Snapchat's AI upgrade

Snapchat is partnering with Google to integrate its Gemini AI system into the platform’s AI elements, replacing or supplementing its previous use of OpenAI's GPT models. According to Snap CEO Evan Spiegel, Google’s Gemini system offers better multimodal capabilities, allowing My AI to process and understand text, audio, images, and video more effectively. This shift is also driven by Google's superior translation features. It's unclear if Snapchat will fully move away from OpenAI, but this change reflects a growing preference for models like Gemini in AI development.