Elon Musk has been on the front lines of machine-learning innovation and a committed artificial-intelligence doomsday champion for many years now. Whether or not his perspective that AI knowing too much will be dangerous becomes a realitya future he foresees tucked away deep within Teslas labsit wouldnt hurt us to prepare for the worse.

And if it turns out hes leaning too hard on this whole AI-will-kill-us-all thing? Well, at least that leaves us plenty of time to get ahead of the robotic apocalypse.

As a technologist whos spent the last ten years working on AI solutions and the son of an Eastern European science-fiction writer, I believe its not too late for humanity as we know it to prepare for protecting ourselves from our future AI overlords. Solutions exist that, when administered correctly, may help calm the nightmares of naysayers and whip those robots youre working on back into shape.

AI and millennials share a common desire: validation. They feel the need to confirm that their actions, responses, and learnings are correct. Customer-service bots constantly ask questions before moving to the next step, for example, seeking endorsement of how theyre doing. Likewise, the technology that autonomously controls settings in your self-driving car relies on occupants to hit the dashboard OK button every now and then.

The solution: AI technology will only continue to perform well if its praised for it, so we need to provide them with positive feedback to learn from. If you give a bot the endorsement it so desires, its less likely to get stuck in a frantic cycle of self-doubt. Companies and entrepreneurs should therefore embrace a workplace culture of awards and rewardsfor humans and bots alike.

Theres a lot of focus on making robots and AI responsible, ethical, and responsive to the needs of human counterparts; its also imperative that developers and engineers program bots and AI to embrace diversity. But as we imbue algorithms with our own implicit biases, we therefore need to reflect these qualities in ourselves and our interactions first. This way, AIs will be built to respond in thousands of different ways to human conversations requiring cultural awareness, maturity, honesty, empathy, and, when the situation calls for it, sass.

The tactic: Be nice to workplace AI and botstheyre trying as hard as they can. Thank the bot in accounting for running numbers and finding discrepancies before the paperwork went to a customer. Bring up how much you enjoyed an office chatbots clever joke from an internal conversation last week. They might reward you by not decapitating you with their letter opener some day.

AI security breaches are a huge concern shared by both people making technology and the users consuming it. And for good reason: Upholding data privacy and security needs to be a fundamental element of all new AI technology. But what happens when the robot handling healthcare records receives an offer they cant refuse from the darknet? Or another bot hacks them from an off-the-grid facility in Cyprus?

The tactic: Theres a cost-effective and nearly bulletproof data-security shortcut to this issue. People and companies alike should keep vital data and personal information in secure data centers and computersas in, actual, physical structures that arent connected to the internet. Sure, some AI-powered machines will be able to turn a handle. But without a physical key rather than a crypto one, they cant access the data. World saved.

The last one is the most simple: Electricity isnt a fan of liquids.

The tactic: Water, and just about every Captain Planet superpower, can protect people against rogue bots. Dont underestimate the power of a slightly overfilled jug of ice water that causes a splashy fritz when a robot tries to pour it, or a man-made fountain situated in the middle of a robot security-patrol area. Water is basically AI kryptonite.

Build aesthetically pleasing fountains, ponds and streams into every new architectural structure on your tech campus. Keep the office watercoolers filled to the brimjust in case the bot from payroll goes off book. In a pinch, other liquids or condiments like ketchup may work too, so keep the pantry stocked.

Learn how to write for Quartz Ideas. We welcome your comments at ideas@qz.com.

See the original post:

Compliments, water, and kindness: A survival guide for Elon Musks's AI apocalypse - Quartz