🖊️ Meta and OpenAI sign agreement to protect global elections

The agreement signifies a unified commitment to develop tools to detect and prevent misleading AI content from influencing public opinion during election time.

They’ll use methods like watermarking or embedding metadata to detect and label political deep fakes and share best practices, so they can respond to deep fakes quickly.

Additionally, they’ll launch widespread public awareness campaigns to educate voters on deceptive content designed to influence opinion.

This follows a movement by the Federal Trade Commission to make AI-voiced robocalls illegal and the EU’s AI Act that requires AI-generated content to be labeled as ‘AI’.

Reddit has reportedly negotiated a content licensing deal with “an unnamed AI company” to allow access to all its posts to train existing Large Language Models LLMs.

The details of this deal—expected to be worth $60M—could change, ahead of its anticipated $5B IPO debut in March, which has been in the works for over three years.

This follows their ‘data cleanse’ which removed years of private chat logs and messages from users' accounts, to ‘prepare for a new chat infrastructure’.

It also comes after reports that Apple and OpenAI have been offering news publishers and media outlets between $5M and $50M for their data.

Anthropic is testing an AI-powered feature that detects when users ask Claude (its AI chatbot) political questions and then redirects them to “authoritative” information sources.

They’ve worked with policymakers, civil society, nongovernmental agencies, and election-specific consultants to develop this, and it will be launched in the coming weeks.

This comes after Anthropic admitted that Claude wasn’t trained frequently enough to provide real-time information about specific elections, so was prone to misinformation.

It also follows the agreement signed by 20 of the biggest tech companies (at the Munich Security Conference) to prevent deep fakes from swaying public election votes.

The founder of SoftBank Group (Masayoshi Son), plans to raise $100B for a new AI chip venture (code-named Izanagi) set to rival NVIDIA, which has been around for 16 years.

The Izanagi venture seeks to complement Arm Holdings’—which Softbank holds a majority stake in—plans to create chip designs for various workloads.

It’s expected that Son will provide $30B with $70B coming from Middle East investors and, if successful, will give it a fifth of the global semiconductor market.

Clubhouse—the live audio app—announced that users can now send texts, and AI will transform them into the sender's custom voice, creating the feeling of being in a real-time chat.

Although it can’t translate laughter, the new feature (which is only available in the US) can recreate custom voices, even when it's only been trained on a few phrases.

This comes after its monthly active users dropped by 93% in January and Apple’s release of ‘Personal Voice’, which uses AI to mimic its user's voices.

A consistent 52% open rate

Best-in-class ad performance

A growing ecosystem of AI products/services