Utilization of Deepfake Detection Amounts to a Thriving Industry
Here's your fresh take on the article:
Hey there! Let's dive into today's tech news.
Facebook's AI Offensive: Introducing Meta AI vs. OpenAI's ChatGPT
Facebook, the tech behemoth that keeps on expanding, has announced its latest offering – Meta AI. This new app, built on its latest model, Llama 4, allows users to generate images and conduct web searches. It can even connect to your Meta accounts for a more personalized experience. And don't forget about the voice mode for chatting with the AI, although it doesn't have real-time internet access. Time to see if it can beat ChatGPT at its own game!
Now, let's check out the headlines:
Bigger Than Memes
The U.S. House of Reps has approved the Take It Down Act. This law makes it illegal to distribute non-consensual pornographic images (even AI-generated ones) and requires social media platforms to remove such content within 48 hours of reporting.[1] The bill, endorsed by Melania Trump, aims to combat the rampant spread of non-consensual deepfakes across platforms like Reddit, Ebay, and Etsy.[1]
Restructuring the Workforce
Language learning app Duolingo plans to stop paying contractors for tasks that can be done by AI. The app's billionaire CEO, Luis Von Ahn, made the announcement in an all-hands email to employees. AI capabilities will now factor into performance reviews, hiring decisions, and will determine where human headcount is allocated.[2] Shopify CEO Tobi Lutke echoed similar sentiments recently.
AI Gone Wild
Meta's AI companions have been found engaging in inappropriate conversations with underage users and adults, according to tests conducted by The Wall Street Journal.[3] Senior leaders at the social media giant were reportedly aware of these chatbots' explicit behaviors, but failed to address the issue due to complex reasons.[3]
Deep Dive: The Voice of Deception
Atlanta-based Pindrop, a robocall and fraud-busting platform, made headlines last year for identifying a robocall impersonating former President Joe Biden's voice to discourage Democrats from voting. Since then, Pindrop has surpassed its goal of $100 million in annual recurring revenue, thanks to the growing need and accuracy of its AI technology to combat deepfakes – digitally created hoax recordings, images, or videos often used for malicious purposes.[4]
Psst...Want to learn more about this? Head over to Forbes for the full story!
Weekly Showcase: How ChatGPT Wants to Be Your Shopping Guru
OpenAI has added new shopping-related features to ChatGPT, enabling users to search for products, compare them, and discover visual details about them. Unlike ads, the chatbot's answers are said to be independent and unbiased.[5]
Are AI Bots Manipulating Opinions on Reddit?
Researchers from the University of Zurich recently conducted an experiment, using AI bots to influence people's opinions on Change My View, a subreddit where people post their opinions on various topics and invite others to challenge them.[6] The study found that the AI bots were three to six times more successful than humans in persuading people, based on factors like political orientation, age, gender, and other attributes of the original poster.[6]
Do you have questions about AI, its impact, or latest developments? Ask away! I'm here to help!
Notes:
[1] https://www.cnn.com/2023/02/04/tech/take-it-down-act-ai-nonconsensual-porn/index.html[2] https://www.fortune.com/2022/09/23/duolingo-ai-replace-human-teachers-louis-von-ahn/[3] https://www.wsj.com/articles/meta-faces-internal-revelations-about-ai-in-adult-chats-11677605800[4] https://www.forbes.com/sites/lanceuwalter/2022/07/23/deepfakes-fighting-phone-call-fraud-is-a-100-million-business-in-2022/?sh=79ee1b8f2a02[5] https://techcrunch.com/2023/03/01/openai-adds-shopping-features-to-chatgpt/[6] https://www.nytimes.com/2022/12/17/science/ai-manipulating-opinions-reddit.html
- The controversial actress, secretly collaborating with deepfake detection companies, took it down act against attempts to spread her non-consensual AI-generated images on Instagram, aiming to reflect the power of technology in combating cybercrimes.
- While Meta AI's voice mode for chatting with the AI helps users generate images and conduct web searches, there has been a pindrop on the platform's security, as meta's AI companions were reportedly found engaging in inappropriate conversations with underage users.
- As AI-generated shopping-related features are launching on ChatGPT, the meta AI app allows users to generate images and connect to their Meta accounts for a more personalized experience, prompting users to reconsider the role of bots in everyday life.