Axon's AI Work: What's Ahead
Artificial intelligence (AI) is an important topic in technology and has the potential to dramatically impact law enforcement. Recent headlines have highlighted these opportunities and the risks that come with them. We'd like to walk you through Axon's roadmap for AI and the steps we are taking to implement this technology responsibly.
Thinking Out Loud
Our role in public safety is based on innovation. We're always thinking ahead. Before we had anyone working on AI, Axon has wondered aloud about how AI could impact customers for years. We have been quite speculative about how AI could be applied to law enforcement in past interviews and in our own communications. Many people, including customers and the public, have asked us 'what if' when it comes to facial recognition and predictive policing, as society as a whole has not thought through these things. That does not mean we are building—or will ever build—those things.
From Speculation to Reality
After the February acquisition of two machine learning research teams, Dextro and Misfit, Axon has the capability to do world-class AI for the first time, and is uniquely positioned to use AI for the public good. This is a responsibility we take seriously. The team has driven the shift from broad speculation to an actual roadmap focused on solving real problems in public safety.
Now that we have a formal effort in place, we are establishing a clean break with previous speculation around AI and focusing on the new roadmap put in place by the Dextro and Misfit teams: one focused on eliminating paperwork and automating tedious back-office workflows such as redaction.
Since its February formation, the team has been focused on redefining our vision and ethical guidelines for the path forward. We fully recognize the complexities and sensitivities around technology in law enforcement, and are committed to getting it right. We take privacy extremely seriously and thus would never release a product that has unaddressed privacy concerns for law enforcement officers or private citizens. To that end, we are finalizing a fully independent, public AI Ethics Board to help guide us. In the coming weeks, we'll announce the members of the Board, comprising established thinkers from civil rights law, data privacy, AI ethics, and the community. We'll then work with them on a set of AI governing principles and publish the results.
Our Focus: Workflows, not Decisions
Even before we assemble the ethics board and define these governing principles, we are committed to accelerating tedious workflows, and to leaving decision-making to the officers and agencies. An example workflow we might address is speeding up supervisor review of footage. An example decision that is outside our scope is inferring criminality.
With this in mind, we are focusing our new AI initiative on one of the most formidable challenges in law enforcement: We're going to eliminate paperwork from the field and triple the time officers can spend serving their communities. To achieve this goal, these are the three areas of investment you can expect from our researchers:
- Automated redaction to speed up the process of sharing footage with the public while protecting the privacy of citizens captured in video.
- Automated transcription to eliminate the heads-down hand-written notes so officers can engage and empathize with the community.
- Automated reporting to populate factual records directly from video and audio, so officers can spend time serving the community.
We're focused on automating tasks that are time consuming for law enforcement officers. Axon doesn't want a world where public servants are robotic. We actually envision the opposite - by removing the burden of manual notes and endless hours of keyboard work, officers can be present, build relationships and be more human. We want our technology to enable more personal interactions to help build safer and stronger communities.
Axon CEO & Founder