I hardly recognize the person I was a year ago. Around this time in 2018, I was working for a little computer vision company that had really big problems. There was a lot of quiet dissent and discontent, hardly any executive transparency, and a whole lot of sudden, unexpected change. I had just been promoted in October — shedding all of the customer-facing responsibilities I had amassed in exchange for managing the company’s data needs.
Things had been changing for a while, but we were frogs slowly boiled. It was hard for me to notice the trend at the time, but looking back it’s harder to see why I stayed as long as I did. A combination of ambition and denial, if I’m honest.
Some personnel changes resulted in new power structures at the company. New sales inquiries from authoritarian regimes were popping into the pipeline, where previously they would have been summarily ignored. It wasn’t long into my promotion before there were proposed tasks asking my team to do things that made us uncomfortable. Things we thought might be illegal, and if they weren’t illegal, then they were things that we thought should be outlawed asap.
I had just been nominated for a chance to work as an ethics officer.
I still believed internal ethics officers could work.
If you’re reading my blog, chances are you know what happened from there. Management refused to agree to my reasonable restrictions on who and where to sell our AI technology. They cancelled plans to have any ethics oversight whatsoever, and I walked away knowing that I had no choice but to find a way to tell the world that autonomous weapons are real, and that privacy is a human right.
That all went down in JANUARY.
It’s crazy to think it’s already been a whole year.
Since last January, I’ve come full circle. I now support two organizations who aim to make AI less harmful, so that the world can benefit from its blessings while avoiding the technology-driven authoritarian nightmares we see happening all over the world. I used to work for a company that sold facial recognition, now I’m part of the coalition trying to burn that intrusive panopticon to the ground. I used to be an ideal tech employee, defending the company at every turn. Now I talk to people who are considering coming forward with secrets, and share with them what the experience of being named in the press was like for me.
A year ago, I wasn’t sure if AI regulation was a good idea.
Now I know that it’s our only hope.
I hardly recognize the person I’ve become: one who yells at 4-star generals for condescending to tech workers; a person who speaks at the UN. I genuinely don’t think I deserve all this attention, but if it means people are talking about killer robots, privacy, and bias more than they were, I’ll continue to speak out. 2020 for me will be a year of making algorithms be less harmful, and attempting to ban or severely restrict surveillance technology (among other applications where AI has no business being used).
But what I’m most grateful for over the last year are the incredible friends I’ve made along the way. I couldn’t possibly name all of them, but I’m so lucky to now be surrounded by some of the kindest, most impactful activists and human rights advocates on the planet. The work of activists like Albert at S.T.O.P, Jack at Tech Inquiry, Peter at ICRAC, and Meredith at AI Now inspired me (before I’d met any of them) to reach out to the ACLU and start this work. To tell the public why AI shouldn’t ever be allowed to take a human life. And now, even in the darkest moments of despair, when I feel like the battle for ethical, well-regulated AI might be extremely uphill, I’m encouraged thinking of how their work has impacted me, and how far I’ve come as a result. Through all of this I have to remember, and to remind others of this important fact… that even though it sometimes seems impossible, and always is difficult working to change hearts and minds, it remains true that…