Welcome to the newest edition of FindBiometrics’ AI update. Here’s the latest big news on the shifting landscape of AI and identity technology:
Google has unveiled new, AI-infused Pixel smartphones that let customers use its Gemini chatbot as an overlay on top of other apps. Its Pixel launch events usually happen later in the year, but the company appears to have been eager to beat Apple to the punch regarding integrating AI into phones. A live demonstration asking Gemini to cross-reference a picture of a concert poster with the phone’s calendar app took three attempts to succeed.
xAI has launched a new AI model that rivals those of OpenAI, Anthropic, and Google. According to Wharton School professor Ethan Mollick, Grok-2 is now one of the “five GPT-4 class models”, alongside GPT-4o, Claude 3.5, Gemini 1.5, and Llama 3.1. The AI model has generated some controversy for a relative lack of guardrails in its content generation, with some users reporting that they could make inappropriately violent and sexual images, and obtain instructions to conduct mass shootings.
Huawei has developed a new AI chip that is comparable to Nvidia’s H100. The Ascend 910C is reportedly being tested by internet and telecommunications companies in China, where Nvidia’s advanced chips are effectively banned thanks to US sanctions.
SoftBank has given up on working with Intel to develop an AI chip that would also rival Nvidia’s solutions. The companies had been negotiating about a collaboration, but Nvidia reportedly couldn’t meet SoftBank’s demands, and the talks fell through before Intel’s recent announcement of thousands of layoffs. SoftBank is now in talks with Taiwan Semiconductor Manufacturing Co.
Canadian VC firm Radical Ventures has raised almost $800 million for a new fund that will focus on later-stage companies. Radical Ventures has previously supported AI startups including Cohere, Covariant, and Untether AI, and has itself received investments from major AI figures including Fei-Fei Li and Geoffrey Hinton.
A law offering redress to the victims of deepfake-generated pornography has unanimously passed through the US Senate. The DEFIANCE (Disrupt Explicit Forged Images And Non-Consensual Edits) Act lets victims seek damages and injunctions for a period of up to 10 years, and lets plaintiffs use pseudonyms to help protect their privacy.
The actors’ union SAG-AFTRA has established a deal allowing actors to license their voice biometrics to advertisers. The agreement was made with Narrativ, an online talent marketplace, and essentially lets advertisers use voice clones of willing actors, who would receive compensation and some degree of control over how their voices are used.
The chatbot’s take: We asked for a little bit of background on a couple of Radical Ventures backers.
–
August 16, 2024 – by Alex Perala
Follow Us