All Hands/Berlin2020/Demos

On Wednesday, January 28 from 16:00 - 18:00 (4pm - 6pm) we'll host 35 demos at the Intercontinental. Here's who and what you can expect (we listed them alphabetically):

Demo Titles (Contact) & Descriptions

Browser Interop Visualisation (contact: James Graham)

A healthy internet requires multiple browser engines that are able to work with all web sites and applications. When content is only accessible in some engines it causes compatibility issues leading users to switch to the browser which works with most sites, almost always the market leader. Left unchecked, this eventually leads to a web entirely controlled by a single vendor; one which is indistinguishable from a proprietary platform.

Even smaller levels of interop difference can lead to the web being hard to develop for; a recent MDN survey showed that 4/5 of the top frustrations for site authors were related to browser incompatibilities. These failings increase the chance that authors will develop sites that only support a subset of engines, or abandon the web for native app development.

This demo will focus on the work to visualise interoperability failures in the web platform. In particular it will showcase the data we get from the web-platform-tests project and how Gecko engineers can use this to find areas where we are incompatible with other implementations before they appear as site compatibility bugs.

Building User Agency via Personal AIs (contact: Richard Whitt)

Advanced AI systems need not be the sole provenance of large, non-fiduciary plaforms motivated solely by their own financial or political gain. Each human being should have the ability to possess her own highly-individualized virtual intelligence, to provide both online and offline support. These computational agents would exist on one’s personal devices, managed for us by trustworthy mediators such as MoCo. This Personal AI, or "PAI," concept has the potential to evolve into an entirely new support system within the paradigm — the essential trusted digital agent.

The demo will show how the PAI interacts directly with the client’s digital “screens, scenes, and unseens” to, among other tasks:

(1) manage and protect their clients’ online and offline data flows, and other digital interactions with third parties; (2) ensure that online recommendation engines are serving relevant information, and not harmful content such as “deep fakes” or addictive videos; (3) challenge the efficacy of financial, healthcare, law enforcement, and other impactful algorithms for bias and other flaws that would harm its client; and (4) actively prevent environmental devices -- smart speakers, facial recognition cameras, biometric sensors -- from needlessly surveilling and extracting our data.

In all such cases, the demo will show how the PAI occupies a critical agential interface, between the ordinary human being it serves, and the vast range of Institutional AIs spanning our digital world. The various core PAI functions and dependencies, and the ways they mediate with the data extraction/influence lifecycle, also will be demonstrated.

Personal AIs are a natural fit for the Mozilla longer-term strategy, and a truly open Web. Simply put, the PAI can become the ultimate locus of human agency in the digital era. By developing and introducing PAIs to consumers, MoCo can become the leading provider of user agency online. Human beings then would be empowered to take greater control over their digital lives, with the PAI acting as the trusted virtual agent, and MoCo acting as the trusted institutional agent. Finally, a healthy Internet puts people first. The PAI can shift the current dominant platforms paradigm, from one where humans are the mere objects of data extraction and influence, to one where humans can become the empowered subjects and authors of their own digital experiences.