5 things about AI you may have missed today: AI deepfakes, Google to invest $2 Bn in Anthropic, more
28.10.2023 - 16:49
/ tech.hindustantimes.com
/ Joe Biden
/ Ai
We have entered the weekend, and yet, artificial intelligence developments are happening at full throttle. In the first incident, a new report has found that due to the possibility of AI-based deepfakes, people online are questioning even genuine videos and photos that are coming from the ground zero in Gaza. In other news, Google has committed to invest $2 billion in the AI company Anthropic, solidifying its relationship with the startup. The move is seen as similar to Microsoft investing in OpenAI. This and more in today's AI roundup. Let us take a closer look.
AI is making it harder for people to trust the images and videos they see online, reports The New York Times. A disturbing image of a charred body initially believed to be a deceased child, circulated on the internet. However, some internet users dismissed it, calling it an AI deepfake. Surprisingly, multiple AI experts found that it was indeed a genuine image. It appears people are not able to trust the content they see online, as a possibility exists that the media might be doctored. This is majorly affecting conveying the ground reports from Gaza.
“Even by the fog of war standards that we are used to, this conflict is particularly messy. The specter of deepfakes is much, much more significant now — it doesn't take tens of thousands, it just takes a few, and then you poison the well and everything becomes suspect,” Hany Farid, a computer science professor at the University of California at Berkeley and an expert in digital forensics, AI and misinformation told NYT.
Google is set to invest up to $2 billion in Anthropic, an AI startup founded by former OpenAI executives, as per a report by CNBC. This commitment involves an initial $500 million cash injection and an additional $1.5 billion to be invested gradually. Anthropic, known for its chatbot Claude 2, a rival to OpenAI's ChatGPT, has gained popularity among companies like Slack, Notion, and Quora. Claude 2 can summarize extensive text, up to around 75,000 words, while ChatGPT can handle about 3,000 words. Anthropic's valuation was $4.1 billion earlier this year.
US President Joe Biden is set to sign an executive order mandating that AI tools undergo security testing before federal agencies can use them, marking a significant step in regulating this technology, a Bloomberg report said. The order utilizes the government's role as a major tech customer to influence changes in AI products from companies like Microsoft and Amazon.
It also instructs the Department of Labor to establish guidelines for preventing discrimination in AI-driven hiring systems, addressing concerns raised by civil rights groups. While US lawmakers are exploring AI limits, Biden's order establishes initial rules as Congress