Is regulation coming?

Everything you need to know in AI this week

September 15, 2023

Well, this is exciting! Welcome to the new Future Tools—the same AI insights you know and love, now delivered to your inbox twice a week. On Wednesdays, you’ll get a roundup of the best AI tools and industry job openings. On Fridays, you’ll get my curation of the week’s most impactful AI news and some personal perspectives. And it all starts now…so hit reply let me know what you think of the new Future Tools!

—Matt

Is AI Regulation Incoming?

Source: AP Photo/Patrick Semansky

For as long as I’ve been covering this space, there’s always been one thing we can count on in AI: a lack of robust regulation (and the job displacement panic, but that’s a newsletter for another time). But the bipartisan blueprint for future AI regulation tentatively proposed by two US senators last week might just change that. What does Congress’ latest effort to catch up with the rapidly evolving AI space mean for builders?

Three things to know:

  1. Licensing: Under the framework, developers working on sophisticated LLMs and so-called “high risk” applications (like facial recognition) would have to obtain a government license. Getting one would require companies to test AI models for potential harm before launch, disclose any times things might go wrong post-launch, and allow third-party audits of of their models.

  2. Liability: Senators Richard Blumenthal and Josh Hawley, who introduced this week’s regulatory framework, doubled down on their previous suggestion that the (in)famous Section 230—which provides legal protections to tech companies for third-party content on their platforms—shouldn't apply to AI applications. This means that people who claim to have been harmed by AI could take AI companies to court should their tools allegedly violate privacy or civil rights.

  3. Transparency: The framework pushes Congress to require AI developers to disclose essential information on training data, accuracy, limitations, and safety to individual and corporate users of their models—and to do it in an accessible way. FYI: Plenty of AI operators have argued that transparency is essential to building robust AI regulation.

Big picture: Congress has been meeting with some of Silicon Valley’s biggest names this month—this AI regulatory blueprint could be pretty influential in those conversations and hearings. But the framework is also likely to receive some pushback…Google and IBM have already argued against the creation of a new AI oversight body.

Want to know more? Elon Musk, Bill Gates, Sundar Pichai, Mark Zuckerberg, and Sam Altman—along with 20 other tech leaders—told lawmakers it’s time for AI regulation. Here are some of the biggest takeaways from that briefing.

Today’s Future Tools brought to you by:
ChatGPT Skool

Become a certified ChatGPT expert and reveal its true power to grow your business:

  • Create high quality SEO content to drive traffic

  • Schedule a month’s worth of social media posts in minutes

  • Use Plugins to make massive productivity gains

  • Learn how to build GPT ChatBots trained on just your data

Come and see why thousands of people from around the world are raving about this comprehensive ChatGPT training program.

🗞️ Other AI News & Articles

Meta Plans a GPT-4 Killer

The AI arms race intensifies: Meta is reportedly cooking up a powerful new LLM set to rival GPT-4, OpenAI’s latest and greatest.

What sets Meta’s LLM apart? Zuck is pushing for his new model to be open source (like Meta’s current model LLaMa 2) and—unlike GPT-4—free to use.

Meta reportedly plans to start training the LLM, which will be focused on helping businesses create their own sophisticated text-based and analytical AI tools, in early 2024.

AI Training Behind Bars

Finnish AI startups are now using a new means to train AI: prisoners.

  • The context: AI companies often pay clickworkers (aka people who perform low-wage, menial online tasks) to label data. But clickworkers are rare in high-wage Finland. So startups like Metroc and Vainu are hiring Finnish prisoners to train their models for, in one case, $1.67 an hour.

  • The controversy: Critics worry that this practice could set a troublesome precedent for future AI prison projects in countries with problematic labor conditions. Because as AI firms seek out previously untapped or underutilized labor forces to keep up with their increasing data labor demands, preventing exploitation will be imperative.

Everything else:

  • Adobe debuted a ton of new AI and 3D features in Adobe Premiere Pro and Adobe After Effects.

  • A federal judge has ruled that AI-generated art can’t be copyrighted.

  • Microsoft has open-sourced EvoDiff, a novel protein-generating AI.

  • Cohesive AI pairs human creativity with AI's brilliance to help you create, polish, and share content effortlessly. End ChatGPT prompt struggles and start crafting world-class content swiftly. Cohesive AI supports you, be it a marketing campaign or your next TikTok. Dive into content creation across text, images, and audio—all in one platform!*

  • Amazon is launching generative AI to help sellers write product descriptions.

  • The Biden administration said it’s secured another round of voluntary safety commitments from eight prominent AI companies, including Nvidia, Palantir, Salesforce, Scale AI, and more.

*This is sponsored advertising content.

📺️ This Week’s Videos to Watch:

More important AI news: If you want to dive deeper into the latest AI news, check my latest YouTube video for everything you might have missed this week. Spoiler alert: It was a busy one.

Cool AI tool: I made an AI song and music video (yes, my album is dropping soon) using free trials. Is this the future of the music industry?

Thought-starter: I thought this interview with Emad Mostaque, CEO of Stability AI, was interesting, especially for high-level observations about the timeline of widespread AI adoption.

And that’s a wrap on the first edition of Future News (the tentative new name for the Friday edition of Future Tools—what do you think?). As always, if you have any feedback/ideas for this newsletter, please get in touch—curious to hear your thoughts on the new format! See you Wednesday for a roundup of the coolest new tools in AI.

—Matt (FutureTools.io)

Interested in working together on a newsletter sponsorship, YouTube integration, or featured listing on the Future Tools website? Get in touch here. I’m pretty picky and look out for the audience above all else but feel free to reach out here and let’s see if we can make something happen.

If you haven’t already, be sure to follow me over on X. I like to share news and AI updates as they happen over there.

And don't forget to check out all of the newest tools we've just added on Future Tools!

You rock! See ya next week. :)

P.S. This newsletter is 100% written by a human. Okay, maybe 96%.