OpenAI CEO Reiterates No User Data Being Used For AI Training — But Not Everyone Is Convinced

Amidst increasing fears around artificial intelligence platforms scrapping user data to train their models, OpenAI’s CEO Sam Altman has reaffirmed the company’s stance.

What Happened: On Tuesday, Altman took to X, formerly known as Twitter, and reiterated the earlier stance stating OpenAI doesn’t use data sent via its API to teach its AI models. 

He further said that they only incorporate user-specified data into future training runs if explicitly requested by customers and only to improve the model in specific ways. 

See Also: Sam Altman Says ‘It’s Hopeless To Compete With’ OpenAI

seeing a lot of confusion about this, so for clarity:

openai never trains on anything ever submitted to the api or uses that data to improve our models in any way.

— Sam Altman (@sama) August 15, 2023

This announcement comes at a time when companies using AI are being closely watched regarding how they use customer data. People are worried about their information being mishandled. 

Here’s how netizens reacted to his post: 

There is no justification for use of private data for training under any circumstances. Your practice of offering conversation history only if users agree to give you access to use and publish their private data is an unreasonable privacy intrusion that ought to be illegal.

— Olav Mitchell Underdal (@omunderdal) August 15, 2023

And.. does the same apply for humans ("end users") using ChatGPT through web interface or app?

Do you train on anything ever submitted to the app or web interface or use that data to improve your models in any way?

Same data ownership principle imo

— Malte✨ (@maltevonmedem) August 15, 2023

hey @sama, you are repeating the same thing over and over again but in my opinion this is not what customers mean when they ask about privacy and OpenAI.

You are talking about "OpenAI privacy" where as customers are concerned about "my LLM app product and data architecture built…

— Anshu Sharma 🌶 (@anshublog) August 15, 2023

Separately on the same day, the ChatGPT parent strongly advocated for integrating AI in content moderation, asserting that it can greatly enhance efficiencies for social media companies.

We've seen great results using GPT-4 for content policy development and content moderation, enabling more consistent labeling, a faster feedback loop for policy refinement, and less involvement from human moderators. Built on top of the GPT-4 API: https://t.co/0HoZjCiStQ pic.twitter.com/lV1Ba7CGaR

— OpenAI (@OpenAI) August 15, 2023

Why It’s Important: For the unversed, earlier this year, Italy prohibited ChatGPT and said OpenAI could only restart the service if certain conditions were met, including providing users with the ability to object to data processing. Since then, ChatGPT’s operation in the country has been restored. 

In April this year, OpenAI launched an “incognito mode” for ChatGPT users, enabling them to decide whether they want the company to save their conversation records and use them for enhancing ChatGPT’s AI capabilities. 

OpenAI has also announced its intention to develop ChatGPT Business — a subscription service tailored for enterprises seeking data control over their end users. 

Earlier this year, industry giants like Samsung Electronics Co. Ltd, Apple Inc., and Alphabet Inc. reportedly clamped down on the internal use of AI-powered chatbots after the South Korean tech giant reportedly faced an issue regarding a confidential information leak. 

Photo Courtesy: Shutterstock.com

Check out more of Benzinga’s Consumer Tech coverage by following this link.

Read Next: After Former Google CEO, Sam Altman Voices Concerns Over AI’s Role In Shaping Future Elections