• Digital Neuron
  • Posts
  • OpenAI Privacy Update, Replit Gets Unicorn Status, US Regulators Warn About AI Bias

OpenAI Privacy Update, Replit Gets Unicorn Status, US Regulators Warn About AI Bias

And A Tool Kit for AI Safety

Author

Digital Neuron
April 27, 2023

Hello and welcome to the third Digital_Neuron newsletter. The best way to keep up-to-date with what is happing in AI, test out new and useful tools, get some extra credit reading and a chance to spread your artistic wings. ONWARD!

Today We Have

  1. OpenAI and user privacy

  2. A AI + Software company gets unicorn status

  3. A tool kit to make AI safer

  4. US regulators taking action on AI bias

Top News Of The Day

  • A little more privacy with OpenAI.

    • OpenAI introduces new features for ChatGPT to offer users better control over their data. These features include data retention settings and data deletion.

      Data retention settings: By default, OpenAI retains user data for 30 days. Users now have the option to shorten or extend the retention period according to their preferences. Organizations can also enforce a custom data retention policy.

      Data deletion: Users can now delete individual data points, like messages or models, directly from the ChatGPT interface. To delete a message, simply click the trash can icon next to it. To delete a model, go to the "Models" tab and select "Delete."

      These updates aim to provide users with more control and transparency regarding their data, ensuring a more secure and customized experience with ChatGPT.

  • Replit gets a billion dollar valuation

    • Replit, an AI-powered coding assistant, has recently achieved unicorn status with a valuation of over $1 billion. The platform significantly improves developer productivity by providing real-time, context-aware code suggestions and error detection. It integrates with popular Integrated Development Environments (IDEs) like Visual Studio Code, PyCharm, and others.

      The company has successfully raised $100 million in a funding round led by prominent investors, including Sequoia Capital and Andreessen Horowitz. Replit plans to use the funds to expand its product offerings, enhance the AI capabilities, and accelerate global expansion.

      Replit's success highlights the increasing demand for AI-driven tools that can streamline and optimize the software development process, making it faster and more efficient for developers.

  • Nvidia releases a toolkit to make text-generating AI ‘safer’

    • NVIDIA has released a toolkit designed to make text-generating AI systems safer and more reliable. The toolkit offers enhanced monitoring and control capabilities to help developers identify and mitigate potential biases or offensive content generated by AI models.

      The toolkit is compatible with popular language models, including OpenAI's GPT series, and can be integrated into existing AI applications. Its features include real-time content filtering, customizable safety thresholds, and detailed insights into model outputs.

      By providing these safety mechanisms, NVIDIA aims to promote responsible AI development and mitigate the risks associated with deploying text-generating AI systems in various industries and applications.

  • Four US regulatory bodies have eye on AI bias

    • U.S. regulators have issued a warning that they already possess the authority to take action against AI bias in financial services. The warning comes amid concerns about the potential for discrimination and unfair treatment in AI-driven decision-making processes, such as lending or insurance underwriting.

      Regulators like the Consumer Financial Protection Bureau (CFPB) and the Office of the Comptroller of the Currency (OCC) are monitoring the use of AI in financial services to ensure compliance with existing anti-discrimination laws, such as the Equal Credit Opportunity Act (ECOA) and the Fair Housing Act (FHA).

      This warning serves as a reminder for financial institutions to proactively identify and mitigate biases in AI systems to avoid regulatory penalties and maintain fair practices.

Useful Tools To Try Out

  • CircleChat lets you use multiple LLMs to solve a issue…or just have them talk to each other and watch hilarity ensue.

  • In HR? Use Covey Scout to streamline candidate selection

  • Insight Copilot get qualitative analysis on a topic by scanning socials

  • Artifact your first AI powered news feed

  • Use BuySmart as a AI purchasing assistant

  • Build AI apps with Promptly

Reads That You Need

Accounts to follow

  • Jonathan Frankle | Chief Scientist at MosaicML

  • Magic Studio | Powered by AI, created by you

  • Sully | Tweets about AI & latest tech trends.

  • Florian S | AI Enthusiast. Software Dev since the age of 6. IT architect for a fortune 500 company. Believer in the good in humanity.

  • Guy Parsons | building things with #AI

  • Sam Altman | OpenAI

Lets start creating “art” together.

  • Send me your best visual prompt and I may run it through Midjourney, Dall-E 2, etc and post it here. For today you will have to settle for something I did a while…