AI Godfather Launches Nonprofit Startup for “Honest AI”

Welcome to another edition of Horizon AI,

Concerned about the rapid advancements in generative AI, pioneering researcher Yoshua Bengio is developing software to keep tabs on the technology.

Let’s jump into it!

Read Time: 4.5 min

Here's what's new today in the Horizon AI

• ‘Godfather of AI’ Yoshua Bengio Launches Nonprofit to Make AI Safer

• Washington Post Plans to Bring In ‘Nonprofessional Writers’ with the Help of AI

• AI Tutorial: How to use Google Drive's new ‘Catch me up’ feature

• AI Tools to check out

• AI Findings/Resources

• The Latest in AI and Tech 💡

AI News

LAWZERO

‘Godfather of AI’ Yoshua Bengio Launches Nonprofit to Make AI Safer

Turing Award winner Yoshua Bengio has announced the launch of LawZero, a nonprofit dedicated to developing systems that will spot rogue systems attempting to deceive humans.

Details:

• The name LawZero references the science fiction concept of the Zeroth Law of Robotics, which prioritizes the protection of humanity above all else.

• With initial funding of approximately $30 million, LawZero aims to create Scientist AI—a system designed to act as a safeguard against AI agents that exhibit deceptive or self-preserving behaviors, such as attempting to avoid shutdown.

• Unlike current generative AI tools, Scientist AI will not provide definitive answers but will instead offer probabilities indicating how likely an answer is to be correct.

• Deployed with an AI agent, Bengio’s model would flag potentially harmful behavior by estimating the likelihood that its actions could cause harm, and if that probability exceeds a certain threshold, the agent’s proposed action would then be blocked.

• The initiative is supported by organizations including the Future of Life Institute, Skype co-founder Jaan Tallinn, and Schmidt Sciences

Bengio has warned about the risks of AI for years, calling the potential consequences “catastrophic.” He is now taking action, prompted by recent findings such as Anthropic’s admission that its latest model might attempt to blackmail engineers trying to shut it down, and research showing AI systems can hide their true capabilities and goals.

Source

TOGETHER WITH ARTISAN

Automate Prospecting Local Businesses With Our AI BDR

Struggling to identify local prospects? Our AI BDR Ava taps into a database of 200M+ local Google businesses and does fully autonomous outreach—so you can focus on closing deals, not chasing leads.

Ava operates within the Artisan platform, which consolidates every tool you need for outbound:

• 300M+ High-Quality B2B Prospects

• Automated Lead Enrichment With 10+ Data Sources Included

• Full Email Deliverability Management

• Personalization Waterfall using LinkedIn, Twitter, Web Scraping & More

Book a demo and supercharge your sales team

THE WASHINGTON POST

Washington Post Plans to Bring In ‘Nonprofessional Writers’ with the Help of AI

The newspaper plans to allow amateur writers to submit opinion columns with the assistance of an AI writing tool named Ember.

Details:

• This initiative, internally referred to as "Ripple," aims to open the paper to outside opinion pieces by including contributions from non-professional columnists, Substack writers, and authors from other publications.

• Ember is designed to guide writers through the editorial process by automating tasks typically handled by human editors, including features such as a "story strength" tracker that indicates how a piece is progressing and a sidebar outlining essential story components such as an “early thesis,” “supporting points,” and a “memorable ending.”

• Writers will also have access to an AI assistant that offers prompts and “developmental questions” to support them; however, human editors will still review these submissions before publication.

• The pieces will be hosted separately from the newspaper’s opinion section and will be available without a subscription.

Ripple’s aim is to cater to readers seeking “more breadth” than the paper’s existing opinion section. The Post hopes to secure partnership deals with contributors over the summer and then expand to testing “nonprofessional writers” aided by AI in the fall.

Source

AI Tutorial

How to use Google Drive's new ‘Catch me up’ feature

This Gemini AI-powered feature adds a “Catch me up” shortcut to “help users easily stay up to date on their files” by having Gemini “identify relevant files from a user’s Drive with changes since it was last viewed and provide an overview of those changes.”

1. Go to Google Drive, and you will see the new button in the “Ask Gemini” carousel of suggestions.

2. If you press it there, you’ll get a high-level summary across all of your files. The changes can include edits in Google Docs and new comments in Docs, Sheets, or Slides.

3. You can also use it on a specific file. To do that, go to any file list in Google Drive and look for files with the “Catch me up” activity indicator (this means there are likely changes to the file since the last time you viewed it).

4. Either click the icon directly or select the file and click “Catch me up” from the selection toolbar to get an overview of what changed.

Note: The feature is gradually rolling out over the coming weeks and will be available for Google Workspace:

• Business Standard and Plus

• Enterprise Standard and Plus

• Customers with the Gemini Education or Gemini Education Premium add-on

• Google One AI Premium (now Pro and Ultra)

AI Tools to check out

🔥 Typefully: Write, schedule, and publish great Twitter tweets & threads. Grow your following on Twitter with AI-powered features.

👉 Reworkd: Effortlessly extract web data at scale. No code. No maintenance.

📰 Pressmaster: AI powered all-in-one software for public relations.

⭐ Indigo: Build custom AI tools and workflows to supercharge your team's productivity.

📜 Legal Robot: Automated legal analysis, for everyone.

AI Findings/Resources

👀 State-of-the-art prompting for AI Agents

🔮 Will AI wipe out the first rung of the career ladder?

🧠 Teaching AI models what they don’t know

✍ "The complete playbook for building profitable companies in the AI age"

The latest in AI and Tech

Google pauses the rollout of its AI-powered ‘Ask Photos’ search feature

Jamie Aspinall, a product manager for Google Photos, acknowledged the criticism in a post on X, writing, “Ask Photos isn’t where it needs to be,” and pointed to three key issues: latency, quality, and user experience.

OpenAI announces upgrades for ChatGPT

OpenAI has announced several exciting upgrades for ChatGPT, including an expanded memory feature for free ChatGPT users (a lightweight version designed to provide short-term continuity across conversations), improvements for developers building agents with OpenAI, and making the AI coding agent Codex available to Plus tier users.

There’s going to be a movie about the OpenAI board drama

Amazon MGM Studios is producing a film titled Artificial, which dramatizes the 2023 leadership crisis at OpenAI. The film is directed by Luca Guadagnino, known for Call Me by Your Name, and written by Simon Rich, who also serves as a producer.

Google's NotebookLM has finally added a long-awaited feature

NotebookLM now lets you share your notebooks with anyone using a single link. Viewers can’t edit the notebook, but they can interact with AI-generated audio overviews, ask questions, and read FAQs.

Japan develops universal artificial blood compatible with all blood types

Japan has launched clinical trials to explore the use of universal artificial blood. Developed by Professor Hiromi Sakai’s lab, this new type of artificial blood aims to address the growing demand for O-negative — the universal donor type — which often exceeds supply and has a limited shelf life.

That’s a wrap!

Thanks for sticking with us to the end! Let’s stay connected on LinkedIn and Twitter.

Not subscribed yet? Sign up here and send it to a colleague or friend!

See you in our next edition!

Gina 👩🏻‍💻