Meta is measuring fairness and bias in computer vision models

⚡ Today’s Highlights

• 📰 News: Meta is measuring fairness and bias in computer vision models + Salesforce and IBM are teaming up to bring AI tools to shared clients

• 💰 Funding: Speak, Superframe, Martian Lawyers Club (MLC), Pirros, and Floworks

• 🌎 Content: Andrew Yang speaks with Peter Diamandis about how AI is changing politics and why Universal Basic Income could help with AI-related job loss

📰 Today's Top Stories

Evaluating the fairness of computer vision models

(10 min read) (Source: Meta AI)

TLDR: Meta has introduced a new AI benchmark called FACET (FAirness in Computer Vision EvaluaTion) to assess the fairness of AI models in classifying and detecting objects and people in photos and videos.

• The FACET dataset consists of 32,000 images of 50,000 people labeled by human annotators. It covers various demographic attributes, occupations, and activities to deeply examine potential biases in AI models.

• Meta encourages the research community to use FACET for scrutinizing the fairness of vision and multimodal AI tasks. This tool aims to help developers understand and mitigate biases in their own AI models.

• Despite releasing FACET, Meta has a questionable history of responsible AI practices. Past incidents include the forced withdrawal of an AI demo for generating racist content and criticisms over its AI ethics team and tools.

The Big Picture: As AI technologies continue to integrate deeply into society and influence decision-making, the focus on developing unbiased, ethical models is intensifying. Meta's new FACET benchmark is a significant step toward transparent fairness evaluation, but it also throws light on ongoing issues like corporate responsibility in the fast-evolving AI landscape.

IBM and Salesforce team up to bring AI tools to their shared clients

(4 min read) (Source: VentureBeat)

TLDR: IBM and Salesforce are partnering to offer Salesforce's AI solutions, including Sales GPT and Salesforce Einstein, along with IBM's industry expertise to mutual clients.

• IBM will leverage its Consulting arm, boasting 160,000 human consultants, and its "IBM Garage" operating model to assist clients in smoothly integrating Salesforce AI technologies into their operations.

• Beyond Salesforce AI solutions, IBM is encouraging clients to adopt its WatsonX enterprise AI platform. WatsonX aims to help businesses unlock and utilize backend data more efficiently.

• IBM's Data Classifier, trained on industry-specific data models, is also recommended to help companies map their internal data, thereby making it more accessible and useful for AI applications.

👀 More Reading

1. The Inventor Behind a Rush of AI Copyright Suits Is Trying to Show His Bot Is Sentient (WIRED)

2. Google’s new pollen mapping tool aims to reduce allergy season suffering (Popular Science)

3. AI will change American elections, but not in the obvious way (The Economist)

4. Companies Will Use Generative AI. But Will They Tell You About It? (The Wall Street Journal)

5. A new AI-generated map of the world’s trees and renewable energy projects could help fight climate change (The Verge)

💰 Funding Alerts

1. Speak, an AI-powered language learning platform that allows users to learn languages by conversing on various topics and receiving instant feedback from an AI language tutor, raised $16 million in a Series B-2 funding round, bringing total funding to $63 million and doubling its valuation in less than a year.

• Speak's language learning platform has expanded from South Korea to more than 20 countries, including Japan, Taiwan, Germany, France, Brazil, and Mexico.

• Round led by Lachy Groom with existing major investors, including the OpenAI Startup Fund, new strategic investors such as Drew Houston and Arash Ferdowsi (founders of Dropbox), and new institutional investors Neo and GSV.

• Funding will be used to accelerate the development of Speak's superhuman AI Tutor, support international expansion, and facilitate hiring to extend availability to new markets.

2. Superframe, an AI-powered software company that focuses on optimizing businesses' go-to-market technology stacks, raised $5 million in seed funding.

• Superframe's AI assistant addresses the accuracy and complexity challenges businesses face with Salesforce and similar tools, enhancing efficiency and reducing resource waste.

• Round led by various angel investors, including data and AI experts, Salesforce consultants, and general operating experts.

• Funding will be used to support Superframe's product development and engineering team expansion. It aims to launch publicly in early 2024.

3. Martian Lawyers Club (MLC), a company focusing on revolutionizing game personalization by leveraging generative AI, raised a $2.2 million pre-seed round.

• MLC aims to transform game experiences from static narratives to interactive conversations, enabling players to shape the game's direction through their actions, providing agency and enhancing personalization.

• Round led by Fly Ventures, with participation from System.One, Amar Shah (co-founder of Wayve and Charm Therapeutics), and Dhyan Ventures.

• Funding will be used to provide an SDK for developers to design interactive game systems that respond dynamically to player inputs.

4. Pirros, a technology platform that optimizes architecture and engineering firms' management of drawing sets, raised a $2 million seed round.

• It uses AI to categorize and catalog design professionals' drawing sets, making them easily searchable and reusable.

• Round led by investors and advisors including Carl Bass, Joseph Walla, Ryan Sutton-Gee, YCombinator, FundersClub, and Twenty Two Ventures.

• Funding will be used to expand the team, enhance the product, and broaden market reach.

5. Floworks, an AI assistant designed to enhance knowledge workers' experience by completing mundane enterprise tasks through natural language, secured a $1.5 million seed investment.

• Floworks aims to transform how knowledge workers interact with software, offering interoperability across multiple applications. It focuses on AI augmentation rather than employee replacement.

• Round led by Y Combinator and Sense AI. Gaingels, Entrepreneur First, and ThinKuvate also participated.

• Funding will be used to expand integrations, targeting sales use cases initially and eventually incorporating other areas of a company.

🌎 Fresh Content

1. In this episode of the Moonshots and Mindsets podcast, Peter Diamandis speaks with Andrew Yang, an American businessman, attorney, politician, and candidate in the 2020 Democratic Party presidential primaries.

Andrew Yang dives into the intricate relationship between AI, politics, and the future of American democracy. Imagine a world where AI doesn't just recommend your next Netflix binge but also manipulates electoral outcomes through deep fakes and fabricated content. Yang goes beyond the headlines to explore how AI could both disrupt and improve our political systems.

How can we adapt to AI's influence on job markets? What's the role of Universal Basic Income in this new landscape? And can technology actually save a political system that many believe is broken?

Andrew Yang sees the potential for AI to serve as a "co-pilot" in our lives, helping us redefine our purpose and education for a more fulfilling future. From the mechanics of electoral reform to the ethics of AI and the feasibility of UBI, this discussion covers a lot of important topics. (YouTube - 1 h 20 mins)

2. In this episode of Freakonomics Radio, guest host Adam Davidson and guests dive deep into the transformative and sometimes unsettling possibilities that AI could bring to our society. The episode tackles a concerning topic that has surrounded AI since its capabilities were revealed to the public: job loss due to AI. With predictions that millions of jobs could vanish, what does this mean for the future of work?

MIT Professor Simon Johnson weighs in on how thoughtful AI development could actually be a force for good, provided we include a diverse group of stakeholders in the decision-making process. The episode also delves into the "alignment problem" in AI—ensuring these advanced systems actually do what we want them to do, without unintended consequences. Examples range from trivial issues like misaligned tax software to more serious concerns like AI systems making confident but incorrect statements that humans cannot understand or rectify.

Check it out! (Spotify - 47 mins)

That's all for today's email! If you want more please follow us at the social channels linked below, or check out our website!

Share our newsletter: If you like our work please share/forward this email with your friends, colleagues, and family. It's the best way to support us!

If this email was forwarded to you please sign up here to continue receiving them.

Want your content, product, jobs, or event featured in our newsletter? Reply to this email with the details, and our team will reach out to you.

Do you use AI for work? Tell us how, and you could be featured in our newsletter!

Check out our website for more resources, including a list of AI investors, products, events, and twitter follows.

For an archive of all our posts, click here.

We'd love to hear from you! You can always leave us comments or feedback by replying to this email!

Powered by AI. Curated and edited by Humans.