Oct 23, 2021
Responsible AI
A lot of people do not trust AI. Sentient machines will eventually wipe out the human race if left unchecked. We’ve all watched Terminator and the birth of SkyNet.
Today, scientists and technologists disrupt at pace. This is progress. It’s great. However, AI requires a bit more thought given the scale and impact it is having on decisions we are asking it to make on behalf of humans. This is important.
A Responsibility AI awakening
I recently listened to a podcast from the lead scientist accredited with the first commercial facial recognition system back in 1994. An amazing technological breakthrough. Now fast forward to 2011 and that very same lead scientist is stepping back to warn us of the impact of what he unleashed within the world. He warns us of the ethics and responsibilities of AI. What changed here?
As it turns out, many different yet interrelated things change. Compute power increased. Data became massively accessible to the masses. Back in 1994 social media wasn’t a thing. So, for facial recognition to be useful, you needed a decent database of digitized faces to look up. That didn’t exist or wasn’t readily accessible. Today, your average coder can scrap our favorite social platforms and build that semi-decent facial database. Next download open-source facial software, connect to widely available cameras, and start identifying people in their local neighbourhood, school, workplace, without consent. Left unchecked, this is a concerning future.
Joseph Atick recognized this and started to lobby for the responsible use of AI four years after his initial breakthrough. We have to get ahead of the technology. Raise awareness and alert people about the ethical impact of technological breakthroughs. What do they mean and how could they potentially change societal norms?
Today, facial recognition systems are the primary mechanism for unlocking our phones. It’s a widely accepted technology gatekeeper to our most personal information. Standard consumer ware.
Do the ‘extra’ work
I’ve designed, built, and successfully deployed many websites of varying shapes, sizes, and scales. I’ve also got it wrong many times. My aha learning moment was that the go-live launch day is urgent, but operational adoption of any web platform is important. Long-term value creation means you never lose sight of tomorrow’s importance whilst dealing with today’s urgent.
So for deploying websites, the finish line is at least 90 days past go-live. Success is the operational adoption of technology by its target users. Execution is merely right to left planning from there. In doing so, we prioritize people and processes ahead of products and platforms. It is even more important to take a similar approach when deploying emerging AI breakthroughs into the world. However, the timelines are longer and the stakes are much higher than websites.
Consider the AI responsibilities that you are accountable for. Give yourself a timeline of at least 9 years after your major technological breakthrough. What do you anticipate the environmental, societal, and human impacts to be? Then educate us. Raise alarm bells. Be honest. Do the extra work. Don’t just blindly revel in the urgency of the breakthrough. That’s a short term rush. Play devil’s advocate. Seek to understand what is truly important.
Congratulations on your breakthrough and thank you for owning the responsibility that comes with it. Giving birth is hard, but raising a child, well, that’s a whole other level of responsibility.
