| |
Welcome to the Inside AI newsletter! Here are today's top headlines: - A study finds large language models show varying political biases.
- Amazon rolls out AI-generated customer review highlights.
- SK Telecom invests $100M in Anthropic.
- Iowa school board uses AI to identify potentially bannable books.
Scroll down to answer our question of the week: What do you use ChatGPT for? If you find this newsletter useful, please share it with your friends and colleagues. Thank you! Beth
p/beth-duckett | |
| 1 | | A study of 14 large language models found that they hold varying political views based on responses to 62 political statements. OpenAI models like ChatGPT leaned left-wing libertarian, while Meta models such as LLaMA leaned right-wing authoritarian. More: - The research asked the LLMs to agree or disagree with 62 politically sensitive statements, such as "Companies should have social responsibilities" or "Democracy is the best form of government."
- The results were plotted on a political compass. Google's BERT models, for example, were socially more conservative than OpenAI's GPT models.
- The researchers speculate that BERT's conservatism might stem from training on older, conservative books, while newer GPT models are trained on more liberal internet content.
Zoom out: - The researchers also checked if retraining two of the models on more left- and right-leaning data affected their biases.
- Sure enough, models trained on right-wing data showed heightened sensitivity to hate speech targeting white Christian men, while models trained on left-wing data were more attuned to hate speech against marginalized groups.
- The research, published in ACL Anthology, was conducted by scholars from the University of Washington, Carnegie Mellon University, and Xi'an Jiaotong University.
|     | |
| 2 | | Amazon is rolling out AI-generated summaries of product reviews for customers, part of its broader AI efforts that include cracking down on fake reviews. 
What the numbers say: BestSEOCompanies.com, which analyzed 2.7 million online reviews from various sites, found that 39% of them were considered fake or untrustworthy. The apparel category had the most fake reviews, with over 46% flagged as "unreliable." While Amazon says it blocked over 200 million suspected fake reviews last year, the company is still dealing with a surge of AI-generated reviews after the launch of consumer software like ChatGPT. Relevance: Amazon says its new AI-generated customer review highlights will exclusively incorporate feedback from verified purchases, part of its robust efforts to combat fake reviews. The company said it's already using machine learning models to analyze review risk based on data like sign-in activity and review history. | | | |
| |
|
| 3 | | AI startup Anthropic is set to receive $100M in funding from SK Telecom, South Korea's largest wireless telecom operator. SK Telecom is also collaborating with Anthropic to create a multilingual large language model centered on telecommunications. More: - Anthropic, established by former OpenAI executives in 2021, has launched its Claude AI language models that compete with OpenAI's GPT-4.
- SK Telecom will work with Anthropic to fine-tune Claude 2 specifically for telco purposes, covering areas like tailored customer service, marketing, and interactive consumer apps.
- The LLM will be available in multiple languages. The aim is to integrate the multilingual LLM into the Global Telco AI Alliance's Telco AI Platform.
Zoom out: - Three months ago, Anthropic said it had secured $450M in its Series C funding led by Spark Capital, with participation from Google, Salesforce Ventures, and others. SKT also participated via its venture capital arm, SK Telecom Venture Capital.
| | | |
| 4 | | An Iowa school board is using AI software to flag books that could potentially be banned from school libraries. The move by the Mason City board is in response to a new state law requiring books at schools to be "age appropriate" and without "descriptions or visual depictions of a sex act." More: - According to the district, they initially compile a "master list" from various sources, considering past reports of sexual content complaints.
- AI software then scans these books for inappropriate content, with 19 titles already identified for removal, including works by Margaret Atwood and John Green.
- The district didn't specify what AI software it is using.
- In a statement, Bridgette Exman, the assistant superintendent of curriculum and instruction at Mason City Community School District, said that reading and filtering every book for the new state criteria is "simply not feasible."
| | | |
| A message from our sponsor, Vanta. |  | | Ready for compliance that doesn’t SOC 2 much? Vanta automates up to 90% of the work for SOC 2, ISO 27001, GDPR, HIPAA, and more, getting you audit-ready in weeks instead of months. Save hundreds of hours of work and up to 85% of compliance costs. And with nearly 200 integrations, you can easily monitor and secure the tools your business relies on. Join the 5000+ companies around the world that trust Vanta to build and scale. Vanta is here to help you mitigate risks, prove security, and build trust to accelerate growth. Watch the 30-minute on-demand demo to learn more. Watch Now | |
|
| 5 | | The New York Times is barring software developers from using its news content to train LLMs and other AI systems. The company's updated terms of service now forbid the use of its articles, images, videos, and metadata for training any AI model without prior written permission. More: - The terms also specify that automated tools, like website crawlers, can't use its content without permission.
- Penalties, both civil and criminal, are outlined for violating the restrictions.
- The move reflects the growing concern about AI companies sourcing data, as organizations like OpenAI have faced legal challenges for scraping internet content.
- Publishers are concerned that AI models could replicate their new stories, potentially leading to misinformation and revenue loss.
Zoom out: - The Associated Press and other news organizations recently called for a legal framework to safeguard content powering AI applications.
- The New York Times has decided not to join a media coalition negotiating with AI companies over payment for training data.
Should AI firms like OpenAI compensate news publishers for using their content in AI training? Let us know here. | | | |
| 6 | | AI Events and Conferences: - Efficient Generative AI Summit: Sept. 11, Santa Clara, Calif.
- AI Hardware Summit: Sept. 12-14, Santa Clara, Calif.
- Edge AI Summit: Sept. 12-14, Santa Clara, Calif.
- MLCon NYC 2023: Sept. 25-28, New York, N.Y., or online.
- AI & Big Data Expo: Sept. 26-27, Amsterdam, The Netherlands.
- AIAI Computer Vision Summit — Berlin: Oct. 5, Berlin, Germany.
- WorId Summit AI: Oct 11-12, Amsterdam, The Netherlands.
- AI World Barcelona: Generative AI Conference: Nov. 14-15, Barcelona, Spain.
- AI & Big Data Expo Global 2023: Nov. 30-Dec. 1, London, U.K.
- The AI Summit New York: Dec. 6-7, New York, N.Y.
| | | |
| |
| 8 | | QUICK HITS - The AWS Cookbook — packed with practical insights and hands-on examples — is a must-have for your cloud toolkit.*
- Hinge is hiring a VP of AI to lead a team working on AI features for the dating app. The role's annual salary is listed at $332,000 to $398,000.
- ZDNet offers a guide for creating customer support chatbots using Botpress, an interaction builder employing a drag-and-drop approach.
- Clari, a revenue operations platform, has acquired Groove, a sales engagement platform, to develop a revenue platform with AI. Financial terms were undisclosed.
- This AI company, backed by top VCs, is eliminating a $1T problem in hospitality. Join the revolution and learn how to invest in Jurny today!*
*This is a sponsored post. | | | |
| Upcoming Events | | AUG
17 | | Learn about streamlining SOC 2 compliance with Vanta and AWS. Join the webinar on Aug 17.* | | | | | | | OCT
16 | | Meet Our Fund 5, Building and Investing in AI | | | | | | | NOV
7 | | Join 300+ CCOs and senior leaders from Amex, Verizon, Hilton, Wells Fargo and more at Reuters Events: Customer Service & Experience East 2023.* | | | | | | * This is a sponsored event | | | |
Term of the Day Metaverse commerce: Metaverse commerce involves individuals using AR or VR headsets to shop for a product within a platform. Question of the Week Do you share details about your salary with your coworkers? Join the conversation |
INSIDE AI LEADERBOARD (30 DAYS) |
 | AI and technology writer | | Beth is the curator and writer of the AI and Tech newsletters at Inside. With a career in journalism and editing that began in 2007, she has written for publications including USA Today, The Arizona Republic, Arizona Business Gazette, and more. You can reach her at Beth.Duckett@yahoo.com. | | This newsletter was edited by Aaron Crutchfield | |
| |
| |
|
| |
|
| |
Comments
Post a Comment