- Work With AI
- Posts
- Google Denies Training Bard with ChatGPT Data
Google Denies Training Bard with ChatGPT Data
March 31, 2023
⚡ Today’s Highlights
📰 News: Google denies training Bard with ChatGPT data and UNESCO calls for action
💰 Funding: DataDome, Agreena, Oscilar, Parloa, and Fixie
🦾 Tools: AI-generated quizzes and flashcards for teachers and custom AI-generated marketing photoshoots
📅 Events: Sign up for The Data Dividend and Healthcare NLP Summit, they’re happening next Tuesday (April 4)!
📰 Today's Top Stories
(3 min read) (Source: The Verge)
TLDR: As reported by The Information yesterday, Google is reportedly trying to improve its AI chatbots and is launching a new initiative called Gemini. Google is now bringing in its DeepMind division to help beat OpenAI and improve its chatbots.
Today, the company has denied the allegations made by a former Google researcher that the company is using data from OpenAI's ChatGPT to train its chatbot, Bard.
The researcher reportedly left Google after warning the company that using the ChatGPT data would violate OpenAI's terms of service and make Bard's answers look too similar.
(5 min read) (Source: Reuters)
TLDR: The Center for Artificial Intelligence and Digital Policy is urging the Federal Trade Commission to halt the release of OpenAI's GPT-4 due to concerns of bias, deception, and privacy risks.
OpenAI's new ChatGPT-4 has been criticized for failing to meet transparency, explainability, and fairness standards, and for exposing private chat histories to other users.
The group's complaint highlights that OpenAI has stopped offering free trials of its GPT-4 due to commercial pressures and concerns over its readiness.
The call for a pause in developing systems more powerful than GPT-4 echoes an open letter signed by Elon Musk, tech leaders, and industry executives citing potential risks to society.
(15 min read) (Source: The Washington Post)
TLDR: Midjourney is a popular AI image generator that has been generating fake visuals of world leaders and other figures. Its lack of uniform standards has made it controversial, especially since some AI tools create fictional scenes involving real people, which can be used for propaganda and harassment.
Midjourney's guidelines disallow adult content and gore and aim to keep its services PG-13 and family-friendly. It is also working on refining AI moderation tools that will review generated images for misconduct, fueled by potentially inflammatory misuse such as the Trump arrest pictures last week.
Midjourney discontinued free trials due to extraordinary demand and trial abuse. Its CEO, David Holz, suggested that nonpaying users were mishandling the technology and its new safeties for abuse didn't seem to be sufficient.
The Big Picture: The impact of AI image generators like Midjourney is concerning, as they can create fake images that can be used for propaganda and harassment. It highlights how artificial intelligence's development has outpaced the evolution of rules for its use.
(5 min read) (Source: UNESCO)
TLDR: The Ethics of Artificial Intelligence Recommendation from UNESCO is a worldwide framework that instructs nations on how to increase the advantages of AI while minimizing the risks it poses.
UNESCO is calling on countries to implement its Recommendation to ensure the ethical use of AI and reduce risks associated with it, including discrimination, disinformation, privacy violations, and environmental and human rights concerns.
💰 Funding Alerts
DataDome, a leading provider of AI-powered online fraud and bot management to companies like Reddit, Rakuten, and AngelList, has raised $42 million in a Series C funding round led by InfraVia Growth and with participation from Elephant, ISAI, and others. DataDome's bot and online fraud solution uses real-time machine learning algorithms to accurately detect and mitigate attacks on mobile apps, websites, and APIs. The funds will be used to expand their commercial rollout and R&D efforts, allowing DataDome to protect more businesses from account takeover, scraping, payment fraud, DDoS, credential stuffing, and other types of bot-driven cyberattacks and fraud.
Agreena is a climate and fintech company that has raised €46 million in its Series B funding round, led by HV Capital and including impact fund AENU and fintech-focused Anthemis. The company’s soil carbon platform uses AI-based monitoring and verification solutions that utilizes satellite imagery, coupled with ground-truth data and machine learning to identify and report on-farm regenerative agriculture practices. They plan to use the funds to expand their geographic footprint and help farms transition towards climate-positive, regenerative farming.
Oscilar, a self-funded company, has created an AI-powered, no-code platform that expedites the quality and speed of risk assessment and decision-making for online transactions. The $20 million self-funding by the co-founders (Neha Narkhede and Sachin Kulkarni) has enabled Oscilar to automate the feedback loop that updates its models and create an AI Risk Decisioning technology that can assess the risk of every online transaction in a few milliseconds.
Parloa has raised €20 million in a Series A funding round led by EQT Ventures, to support its expansion into the global customer service and technology markets and increase product capabilities. Parloa partners with OpenAI’s language models via Microsoft Azure to offer an AI-based customer service platform that uses natural language to solve customer inquiries and can conduct thousands of automated customer calls and chats simultaneously.
Fixie, a cloud-based platform that uses NLP and AI to automate business processes, answer data questions, and integrate with various tools and APIs, has secured $17 million in seed funding led by Redpoint. Other investors in the round include Zetta Venture Partners, SignalFire, Bloomberg Beta, and Kearny Jackson, as well as angel investors from tech giants like Google, Amazon, and Apple. The company plans to use the funds to grow its team and scale its platform, which is now available for anyone to integrate with their own applications, data, and tools via the Fixie Developer Preview. You can check out their open job postings here.
🦾 Trending Tools
Kittl AI integrates AI image generation into the Kittl design process, offering seamless workflow and unmatched speed. Its browser-based vector design platform enables text-to-vector, text-to-image, background removal, and mockup visualization. The first design AI tool with vector editing.
Quiz Wizard saves teachers hours with AI-generated quizzes and flashcards. Generate content in seconds from a topic or document and export to live quiz or memorization platforms.
Visualize is an AI-powered design tool that turns your ideas into reality in just seconds. With 100+ customizable styles, you can create prototypes, visualize designs, and get inspired in no time. Perfect for hobbyists and professionals, it's easy to use and a whole lot of fun.
Claid Scene Creation is a customizable generative AI tool that transforms product images into photoshoots for marketing, social media, and ads. It offers 100+ customizable backgrounds, realistic scenes, and branding consistency.
Spline AI is launching an alpha of their 3D design collaboration tool that uses AI to generate objects, scenes, and animations in real time. It simplifies and speeds up the design workflow, allowing users to focus on creative goals.
Wonderchat enables users to create AI chatbots in 5 minutes, embedding them on websites for self-serve customer support. Features include chat history logs, learning from mistakes, and API access, with use cases including sales and customer support.
Sttabot is a no-code tool that transforms prompts into a fully functioning AI application within 60 seconds, allowing anyone to create their own custom AI solution. Join the waitlist to try it for free.
👀 More Reading
DoD Chief Digital and Artificial Intelligence Office Launches Hack the Pentagon Website (Department of Defense)
How PNNL Is Using Machine Learning To Detect Nuclear Threats Quicker and Easier (Pacific Northwest National Laboratory)
BuzzFeed Is Quietly Publishing Whole AI-Generated Articles, Not Just Quizzes (Futurism)
NASA-enabled AI Predictions May Give Time to Prepare for Solar Storms (NASA)
What the AI-generated image of Pope Francis means for the imagination (Vox)
Vinod Khosla on how AI will ‘free humanity from the need to work’ (Semafor)
Fighting The AI Tide: Exercise In Futility, Or In Raising Long-Term Awareness? (Forbes)
🌎 Fresh Content
1. Lex Fridman’s latest conversation is with Eliezer Yudkowsky, co-founder and research fellow at the Machine Intelligence Research Institute (MIRI), who is best known for popularizing the idea of friendly artificial intelligence and his strong stances against the acceleration of AI. The conversation was extremely interesting and enlightening, especially with talks of pausing AI progress flying around. (YouTube)
2. Ethan Mollick had an awesome Turing-test-off between Claude, Bing, Bard, and GPT-4 and the results were very interesting. It seems like the chatbots are still pretty confused about identifying human behaviour. (Twitter)
Its the great AI Turing-Test-Off!
I asked Claude, Bing, Bard & GPT-4 to ask a question, and guess which answer was written by me, and which was written by one of the other AI models
Bing & Claude both thought GPT-4 was human. GPT-4 thought Bard was. No one thought I was human. https://
— Ethan Mollick (@emollick)
4:08 AM • Mar 30, 2023
🧠 Resources
1. Discover how to expand the capabilities of ChatGPT with these extensions and create your own ChatGPT app in this video. Explore the benefits of ChatGPT Writer, Merlin, ChatGPT for search engines, and more, to enhance your AI experience.
2. Here are some prompts for you to try out. Try them with a generative AI tool of your choice (ChatGPT, Bing Chat, Claude, etc.) to see how AI can make your life easier, more efficient, or just a little more fun.
Comedy: I want you to act as a stand-up comedian. I will provide you with some topics related to current events and you will use your wit, creativity, and observational skills to create a routine based on those topics. You should also be sure to incorporate personal anecdotes or experiences into the routine in order to make it more relatable and engaging for the audience. My first request is "I want a humorous take on politics." (In this case, politics is just an example. Try out a variety of topics and see what kind of material it can come up with)
Persuasive Copywriting: I'm looking for a [type of text] that will speak directly to the needs and pain points of my [ideal customer persona] and persuade them to take [desired action] with a sense of urgency and strong offer.
Marketing: Please write a marketing campaign outline that leverages the Pareto Principle to identify the most important [product/service features] for [ideal customer persona] and focuses on maximizing the impact of these features. Consider how to prioritize the remaining [20%/80%] of features in a way that adds value to the customer experience.
Learning Philosophy: I want you to act as a philosophy teacher. I will provide some topics related to the study of philosophy, and it will be your job to explain these concepts in an easy-to-understand manner. This could include providing examples, posing questions or breaking down complex ideas into smaller pieces that are easier to comprehend. My first request is "I need help understanding how different philosophical theories can be applied in everyday life."
📚 Research
📅 Upcoming Events
Introducing The data dividend: harnessing the power of people, processes and technology to unlock value from data (April 4, 2023 / Boston, MA, USA) a forum that brings together leaders in government and industry to discuss how to establish the strategy, culture and skills needed to foster a data-driven enterprise and cultivate new ways of thinking and working. They’ll explore questions like:
What is top of mind for data and analytics leaders in 2023? What key challenges are they facing?
How is the regulation of data and AI evolving, and how will it affect your industry and organisation to bring new opportunities and challenges?
Can we trust decisions and recommendations made by AI? How can tech leaders mitigate risk, bias and harm?
Healthcare NLP Summit (April 4-5, 2023 / Virtual) is a two day event consisting of immersive, industry-focused content including over 30 technical sessions, focusing on NLP applications in healthcare and life sciences. Join live Q&A sessions with the speakers, connect with others through networking features, and access all content on-demand after the event. The second week will feature beginner to advanced live training workshops and certifications. Learn, share, and apply best practices for putting AI to good use! Register for this free online event today!
💼 Jobs
New York, NY, USA (Hybrid)
$172,000 - $208,000 USD
Bellevue, WA, USA (Hybrid)
$175,000 - $220,000 USD
Bungie: Data Engineering Architect (Machine Learning)
USA (Hybrid/Remote)
$200,000 - $235,000 USD
Amazon: 2023 Fall Applied Science Internship - Machine Learning - United States
Seattle, WA, USA
$55.34 - $102.31/hour
Remote
Widlcard, discuss your strengths and come to a pay that suits your role
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!
How'd you like today's email? |
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!