| |
| Hey Insiders! Welcome to today's newsletter for Tuesday, July 11. Here are our top stories: - Anthropic launches Claude 2 to the public for testing.
- Detectors used to identify AI-generated work can discriminate against non-native English speakers.
- VC firm Sapphire Ventures plan to invest $1B+ in AI startups.
- China could require licenses for companies to release generative AI models.
Beth
p/beth-duckett | |
| 1 | | Anthropic, a startup founded by former OpenAI research executives, has launched its latest AI chatbot to the public. Users in the U.S. and U.K. can now try out the beta version of Claude 2 at claude.ai. More: - Anthropic is also opening up the Claude 2 API to developers and businesses at the same cost as its predecessor, Claude 1.3.
- Claude 2 offers improved capabilities, including more "harmless responses," better performance on tasks like multiple-choice exams, and the ability to summarize up to 75,000 words.
- It scored 76.5% on the Bar exam's multiple choice section, up from the prior version's 73%.
- Anthropic's AI models have also undergone testing by companies like Slack, Notion, and Quora.
- The generative AI company has amassed a waitlist of over 350,000 individuals seeking access to Claude's API and consumer products.
Zoom out: - Anthropic has raised $750M in funding since March when it had a pre-investment valuation of $4.1B.
- Crunchbase estimates that the startup, backed by Google, Zoom Ventures, and others, has raised outside investments totaling $1.45B, making it the third best-funded generative AI startup.
|     | |
| 2 | | What happened: The Organization for Economic Co-operation and Development (OECD) has warned that over 25% of jobs in its member countries could be easily automated by AI in the future. What the numbers say: According to the OECD's 2023 Employment Outlook, jobs with the highest automation risk make up an average of 27% of the labor force, with Eastern European countries being especially vulnerable. Jobs were considered high risk if 25 out of 100 of their skills were easily automatable, according to AI experts. What it means: Automation will take over some jobs. Three out of five workers fear losing their jobs to AI within the next decade, according to another OECD survey from 2022. However, the same survey revealed that two-thirds of workers already using AI said automation had made their jobs less dangerous or monotonous. Quotable: OECD Secretary General Mathias Cormann said governments have a responsibility to create policies and help employees prepare for changes coming due to AI. Meanwhile, measures like minimum wages, collective bargaining, and government regulations could help mitigate the negative effects of AI on workers and wages, according to the OECD. 
| | | |
| A message from our sponsor, Fingerprint. |  | | Is your device intelligence model keeping up with the speed of fraud? Low-accuracy identification can prevent your fraud model from making good decisions by failing to identify repeat fraudsters when they return. Fingerprint's device intelligence platform can recognize returning visitors with 99.5% accuracy by connecting multiple signals. In just the last month, Fingerprint identified 4.3 million bad bots, 14 million users in incognito mode and 21 million users browsing with a VPN. Find out how | |
|
| 3 | | Researchers have found that detectors used to identify AI-generated work, such as essays and job applications, can unfairly discriminate against non-native English speakers. The study tested seven popular AI detectors on over 90 essays written by non-native English speakers for the Test of English as a Foreign Language (TOEFL). More: - In the results, over 60% of the TOEFL essays were wrongly flagged as AI-generated.
- When essays written by native English-speaking eighth graders in the U.S. were tested, over 90% of them were correctly identified as human-written by the same tools.
- According to the researchers, the bias is likely rooted in the concept of "text perplexity," which evaluates the predictability of words in a sentence.
- Low perplexity is typically associated with AI-generated text, and non-native speakers, who often rely on common words, are more likely to exhibit this predictability.
- The study emphasizes the potential harm this misclassification can cause, including discrimination against non-native English speakers in academic and job settings.
| | | |
| 4 | | Sapphire Ventures, a venture capital firm, is committing over $1B to invest in AI enterprise startups, including companies developing generative AI. The Austin-based firm has a history of backing AI startups and has already invested $2.4B in 60 AI-focused companies, including DataRobot and Clari. More: - Sapphire's AI investment strategy will focus on backing B2B software companies that leverage data to enhance prediction capabilities, making AI easily accessible.
- The majority of the funds will be allocated to direct investments in AI startups, with some going to early-stage AI-focused venture funds.
- The firm also plans to create an "AI Community" for portfolio company executives and incorporate AI tools across its internal workflows and job functions.
Zoom out: - Sapphire Ventures joins other companies, such as Salesforce Ventures, Workday, and Accenture, in making substantial investments in AI ventures.
- According to PitchBook, AI startups have defied the recent decline in VC funding, raising an impressive $15.5B this year. The median post-money valuation has also risen by 109.8% compared to last year.
| | | |
| 5 | | China's internet watchdog is planning to introduce a licensing requirement for AI companies before they can release generative AI models, according to The Financial Times. The aim is to strike a balance between promoting AI and technological advancements and maintaining control over content, sources said. More: - According to FT, China's government wants to regulate generative AI before it becomes widely adopted.
- Companies like Baidu and Alibaba have already been working with regulators to ensure compliance with the regulations, which were first proposed back in April.
- The draft rules state that AI content should align with "core socialist values" and avoid undermining state power or national unity.
- The strict requirements could present challenges for companies in developing and improving large language models, potentially leading to cautiousness in the industry.
- Finalization of the rules is expected to occur as early as this month.
| | | |
| 6 | | New job postings in AI: - Staff Software Engineer, Machine Learning Systems, Accelerator Performance
- Google | Sunnyvale, California
- ML/CV Research Engineer
- Orchard Robotics | Bellevue, Washington
- Artificial Intelligence Machine Learning Specialist
- Booz Allen Hamilton | McLean, Virginia
- AIML — Sr ML Engineer, Data and ML Innovation
- Apple | Cupertino, California
- Research Scientist — AI Compilers
- Nvidia | Redmond, Washington; Santa Clara, California
- Research Engineer — AI — Machine Learning — Vernova Research
- GE Corporate | Niskayuna, New York
- VP, Artificial Intelligence & Machine Learning
- Fortrea | Remote (U.S. and U.K.)
| | | |
| 7 | | QUICK HITS - Combine traditional project management with your docs, knowledge base, and AI.*
- In a blog post, Microsoft co-founder Bill Gates says AI-generated misinformation and deepfakes have the potential to "undermine elections and democracy," making it easier to tilt elections in a particular candidate's favor.
- University of Chicago researchers developed Glaze, a program that shields art from AI training by distorting the input data. Released in March, the freeware now has 890,000+ downloads.
- Vellum.ai, a startup offering a developer platform for building LLM apps, has closed on a $5M seed round that included participation from Rebel Fund, Eastlink Capital, Pioneer Fund, and Y Combinator.
- Future-proof your skills and stay ahead with Brilliant. From AI to math and data science, their interactive lessons help you get ahead. Try it free today.*
*This is a sponsored post. | | | |
| Upcoming Events | | JUL
12 | | Monthly Meditation guided by Nicholas Whitaker | | | | | | | JUL
13 | | #ThankfulThursdays Wellness Community Challenge | | | | | | | OCT
16 | | Meet Our Fund 5, Building and Investing in AI | | | | | | * This is a sponsored event | | | |
Term of the Day Affiliate marketing: Affiliate marketing is a type of performance-based marketing where a company partners with individuals who promote their products for a commission. Question of the Week Do you think it's ever acceptable to bring politics into the workplace? Join the conversation |
INSIDE AI LEADERBOARD (7 DAYS) |
 | AI and technology writer | | Beth is a writer and editor at Inside.com covering artificial intelligence and daily technology news. She has written the Inside AI newsletter since 2019. You can reach her at Beth.Duckett@inside.com. | | This newsletter was edited by Aaron Crutchfield | |
| |
 | Identify up to 99.5% of returning visitors on your platform, mitigate fraud, and improve user experience with Fingerprint. | |
|
| |
Comments
Post a Comment