So @bemyeyes is opting us all into AI training models and we have to opt out. *sigh* I guess I'll just delete my account, then, since this isn't opt in! The below is an excerpt from their latest newsletter sent last night.
Tomorrow we will be making an important announcement, and we wanted you as a user of our products to know first. The announcement is about the fast-paced developments in the artificial intelligence (AI) sector, along with our own role and responsibilities in it:
1. We intend to provide video data to organizations to train AI models to be more inclusive and representative of the blind and low vision community
2. We are strengthening our data and privacy policy - and making it more restrictive - to ensure you have even greater control over if and how your data is used
We are taking these steps to ensure our users, and the blind and low-vision community generally, are not left out of the innovative power of AI, and that AI models are trained using data that reflects the real, relevant, and lived experiences of people who are blind or have low vision.
The 26th of July this year is the 34th anniversary of the signing of the Americans with Disabilities Act (ADA), the United States’ strongest measure protecting the civil rights of people with disabilities. With equivalent acts around the world such as the European Accessibility Act (EAA), there has perhaps never been a stronger focus on ensuring equality and inclusion for all. However, with the exponential rise of AI technologies beginning to impact every element of our lives, there has also never been a greater risk of undoing all the great work that has been done to-date.
That’s because AI is developing at a breakneck speed, and we are already starting to see examples of biased characterizations of blindness and even offensive language being surfaced through AI, as well as less than inclusive responses to otherwise reasonable questions. There is no suggestion that this is deliberate. Blind people have simply not had the opportunity to inject their own experience into the extremely fast-paced development of AI models. So the potential for AI-generated bias and exclusion is very real, and completely contrary to the principles of the ADA, the EEA, and our own Be My Eyes’ mission.