Consider this: AI and Internet Health

Science fiction depictions of artificial intelligence — like the terminator — inform our societal perceptions. In turn, we usually perceive AI in one of two ways: an existential threat, or a panacea.

This simplistic narrative is not only wrong — it’s dangerous. Both sides of the popular ‘AI good’ vs. ‘AI bad’ narrative leave humanity and society out of the decision making loop. They suggest we have no opportunity to shape where AI goes and what it is used for. We just need to wait around for ‘the future’ to accept our fate.

The thing is, the future is already here. Many of the technologies we commonly throw under the umbrella ‘AI’ have already entered the mainstream of computing and everyday online life. More importantly, we can decide where to take these technologies as we are creating them. We can also start to map out and respond to the consequences of using these technologies as we roll them out.

It’s with this in mind that some of us at Mozilla have been asking: What role can Mozilla and the community that it is part of play in pushing AI in a better direction for humanity? What approaches to AI will make it more likely that we create a healthier digital world?

The one thing we have decided so far is: Mozilla is mostly likely to have an impact if it focuses its efforts on the consumer tech space, including the development and use of AI by the large tech platforms. From there, we are working to narrow down areas where we might have an impact. Our leading options are:

  • Accountability. Citizens and governments successfully hold companies to account for AI systems that make discriminatory decisions, abuse people’s data, or harm people. This sparks a race to the top by companies that want to be seen as trustworthy in the eyes of consumers.
  • Agency. People understand when machines are making decisions for them, and can shape, question or opt out from these decisions. They feel empowered and safe.
  • Rights. AI is developed, used, and constrained in ways that protect people’s rights to privacy, free expression, and non-discrimination, especially in relation to vulnerable populations.
  • Open Source. The number of open source AI tools, models, and datasets grows, decentralizing core AI innovation and opening up space for more players outside the big tech companies.

We will pick from these options — or create a hybrid version — to help us prioritize the resources we put into things like the Internet Health Report, fellowships and awards and advocacy campaigns. All of Mozilla’s movement building activities are designed to support people and organizations who share our commitment to creating a healthier internet. We’re already seeing an increased momentum on AI issues across the movement. Our aim here is to figure out where we can best help to grow this momentum.

In order to do this, we want your feedback. I invite you to visit a site called Consider.It where we have posted these options as well as some additional context. If you care about these issues — and, especially, if you are part of the community Mozilla works with, please take some time to tell us what you think. In particular, we want to know what you like, what you don’t like, where we can make adjustments in these outcomes. Don’t hold back. We’re inviting open feedback until May 26th.

Once we get all of your feedback, we will hone our options and bring them into discussions with Mozilla staff and board members in late June and early July.

PS. As we’ve been developing these options, a number of people have asked: ‘Why use the term AI? Isn’t it over hyped?’ Yes, it is. And that is part of the reason to use it. We’d originally talked about a focus on ‘better machine decision making’ and found that we were out of sync with how others were framing the issues. Our goal is to shape the debate — and the use of these technologies. For that, we had to go with the term that is more widely used. We still have work to do with our colleagues on the technology side of Mozilla to better define the scope and definitions that bound our work here. We will share more on this over the coming months.

Comments

  1. Zafar Satyavan replied on | Reply

    Mark, what I feel is that Ai/Vr are acronyms that suggest a certain fascination for machines without properly acknowledging the human factor. It will always be easier to make a human being to think like a machine than the other way round. I am a Human Being and I like the idea that technology can AUGMENT the reality I perceive through my senses.. that works for me!!

Add a Comment