• Work With AI
  • Posts
  • Microsoft and Adobe Team Up to Fight Deepfakes

Microsoft and Adobe Team Up to Fight Deepfakes

1-30-2023

⚡ Today’s Highlights

  • 📰 News: Microsoft/Adobe team up to fight deepfakes and Meta uses AI to get around Apple's new privacy rules

  • 💰 Funding: Anthropic and Hawk AI

  • 🦾 Tools: Conversational search engines

  • 📅 Events: The RE•WORK Virtual AI Summit

📰 Today's Top Stories

(10 min read) (Source: cbsnews.com)

Microsoft and Adobe have teamed up to tackle the problem of deepfakes, AI-generated videos that threaten privacy, the integrity of news, and the democratic process. Instead of trying to spot fake content, they've created Content Credentials, a tool that verifies the authenticity of images by tracing their origin and edits. 900 companies across the photo and video industry have agreed to display the Content Credentials button. Eric Horvitz, Microsoft's Chief Scientific Officer, believes this has a chance to make a real impact in addressing the challenges of deepfakes.

Deepfakes are becoming a more prominent problem with the growing popularity of generative AI. They present several ethical and legal issues related to people’s privacy and the growing problem of dis- and misinformation. Privacy laws and education will also play a role in how legislation is created to combat this trend. Hopefully, AI can help solve this problem as much as it’s exacerbating it.

(15 min read) (Source: wsj.com)

Meta, the parent company of Facebook and Instagram, is seeing a path to recovery after a difficult year of heavily investing in AI tools. They’ve been used to improve ad-targeting systems, allowing the company to make better predictions with less data. The company has also shifted to forms of advertising that are less dependent on harvesting user data to overcome restrictions imposed by Apple's privacy changes. AI tools have helped boost Meta's short-form videos (Reels) and improve their relevance to compete with YouTube and TikTok. Despite the additional strain on computing power and servers, Meta is employing AI more than ever before to use its data more efficiently so that it can detect deeper correlations in user behaviour and improve its recommendations.

With Meta joining the landscape of tech giants employing new AI tools in their operations, there are bound to be rapid advancements in how data is collected and used. It will become an important tool for Meta to use in getting around Apple’s innovative new privacy rules that have made it harder for companies to gather user data.

(8 min read) (Source: vice.com)

Recent videos have surfaced online, generated by AI, expressing support for Burkina Faso's military junta which took control of the country through a coup in September. The videos were created using the Synthesia AI video generator. A professor of digital culture in the UK stated that she did not believe that the videos would be convincing to the public. The CEO of Synthesia confirmed that the videos violated the company's terms of service and the user was quickly banned. Check out the videos here, the production quality is… questionable. 

The misuse of AI is becoming more frequent and many important figures in the AI community, including the CEO of OpenAI and the CEO of Synthesia, are pointing out how difficult moderation can be in these cases. Cases that don’t use explicit or hateful language but may share harmful ideologies are incredibly difficult to detect and systems are being improved every day to prevent them from happening.

(7 min read) (Source: techcrunch.com)

SoundCloud rappers will soon be levelling up in a major way. Google has created an AI musician named MusicLM that can compose songs in various genres based on text descriptions. It's trained on 280,000 hours of music, but unfortunately, it's not quite ready for widespread use. The system has several limitations that need to be addressed like distorted quality, poor vocals, and a tendency to copy copyrighted material (like most other generative AI). Google will be keeping this project under wraps for the time being while they iron out the ethical and legal challenges posed by these new tools. You can read further into the report released by Google here.

Generative AI tools have been the source of many legal problems recently, particularly involving copyright and the stealing of intellectual property. Do you think Google is making the right move by delaying the release of its new tech?

💰 Funding Alerts

  1. Hawk AI, a company which helps “financial institutions detect financial crime more effectively and efficiently using AI to enhance rules and find anomalies”, announced $17M in Series B financing “to accelerate product development and global expansion.”

  2. Anthropic is an AI safety and research company that’s working to build reliable, interpretable, and steerable AI systems. Building on the approximately $700 million raised since the founding of the company in 2021, this round of funding is reportedly close to $300 million, bringing the valuation of the company to $5 billion.

🦾 Today's Top Tools

🌎 Today's Top Content

1. We’re just as confused as the guy in the picture. (Twitter)

2. Robot dogs, an AI’s best friend. (Twitter)

4. Report Says AI Could Potentially Replace 85 Million Jobs Worldwide By 2025. (Reddit)

5. Check out this discussion about reaching technological maturity quickly. (Reddit)

👀 Further Reading:

🧠 Resources

  1. While we wait for Google to release MusicLM, why not check out 2023’s AI models for audio generation?

 📅 Upcoming Events

The RE•WORK Virtual AI Summit (January 30-31, 2023), a chance to hear the latest technology advancements, practical examples of how to apply AI to solve challenges in cross-industry settings, business and society, and delve deeper into the work of leading AI experts in a series of presentations, panel discussions, interviews and fireside chats.

 💼 Jobs

Powered by AI. Edited and curated by Humans.

We'd love to hear from you! Leave us comments or feedback by replying to this email!