Privacy

Using posts to train AI

By Chiedza, June 12 2024

Meta, Facebook and Instagram’s parent company plans on using public Instagram and Facebook posts to train AI from June 26. This includes posts, photos, and captions, but not private messages. The company will notify users in the UK and EU, making it possible for them to opt out by completing a form.

However, even if someone opts out, their data could still be used if they appear in other users' public posts. Meta emphasized that the only information that would be used is from posts not comments, for people over the age of 18. The main goal for this is to develop and improve AI research and product development.

Continue reading: 
Meta will soon use your public posts to train its AI. Can you prevent it?
AI
Social media
Privacy

ChatGPT is a blurry JPEG of the web

By Matthew, May 01 2024

OpenAI’s chatbot offers paraphrases, whereas Google offers quotes. Which do we prefer?

The article explores the capabilities and limitations of ChatGPT, an AI language model developed by OpenAI, particularly in its ability to understand and generate human-like text. It discusses how ChatGPT is trained on vast amounts of text data from the internet, enabling it to mimic human conversation and provide contextually relevant responses.

However, the article also highlights the model's shortcomings, such as occasional inaccuracies, lack of understanding nuanced contexts, and potential biases inherited from the data it was trained on. Additionally, it delves into the ethical considerations surrounding AI language models like ChatGPT, including concerns about misinformation, manipulation, and privacy. Overall, the article provides a nuanced examination of ChatGPT's capabilities and the broader implications of AI language models for society.

Continue reading: 
ChatGPT is a Blurry JPEG of the Web
AI
Misinformation
Ethics
Privacy