AI Update: Fear Makes Strange Bedfellows

Welcome to the latest edition of FindBiometrics’ AI update. Here’s the latest big news on the shifting landscape of AI and identity technology:

The United Kingdom’s AI Safety summit has produced a joint statement signed by strange bedfellows. The United States, China, Saudi Arabia, Japan, and the European Union were among signatories acknowledging the need for cooperation in the face of emerging artificial intelligence technologies that have the potential to cause “catastrophic harm”.

The statement arrived after an executive order from the White House that laid out AI safety rules. The order includes a provision allowing the executive branch to use the Defense Production Act to compel companies to share AI safety testing information if their models pose serious security, health, or economic risks. The order also requires the Department of Commerce to draft guidance on watermarking AI-generated content.

Amazon Web Services has launched a new service that lets companies reserve access to GPUs on certain dates and times, thanks to a consumption model called “Amazon Elastic Compute Cloud (EC2) Capacity Blocks for ML”. Customers will specify the duration of their session and the size of the GPU resources needed. There are no refunds.

Paris-based AI startup Mistral is aiming to raise $300 million in a new investment round, four months after raking in $113 million in seed funding. The company describes itself as the “OpenAI of Europe”. It’s not clear which venture firms may be in talks; Lightspeed Venture Partners led the previous round.

Google has committed to investing $2 billion in Anthropic. The funding will comprise an upfront investment of $500 million in the OpenAI rival, with Google agreeing to pour in the rest over time. Amazon had previously said it would invest up to $4 billion in Anthropic.

AMD expects a significant uptick in revenues from its GPU chips. In the company’s earnings update this week, CEO Lisa Su said GPU revenues are expected to come in at $400 million in the fourth quarter of this year, and to exceed $2 billion over the course of 2024. Its new M1300A and M1300X GPUs are expected to read volume production in Q4 of this year.

A deepfake video of Bella Hadid expressing support for Israel in its conflict with Hamas has gone viral. The clip of the Dutch Palestinian model was posted on the X social media platform by an Israeli user, and eventually received a ‘community note’ flagging it as AI-generated.

In Canada, a resident of British Columbia says that videos depicting him criticizing the Prime Minister are deepfakes. Liu Xin is in fact a vocal critic of the Chinese Communist Party, and he says the latter is trying to tarnish his name with AI-generated “spamouflage” videos.

The chatbot’s take: We asked ChatGPT about whether its parent company should be worried about the competition in Paris.

–

November 2, 2023 – by Alex Perala