Humanitas Ex Machina

Humanity in the bot age

Ariel Jalali
Sep 16, 2016 · 11 min read

Point A to point B

I did a crazy thing two years ago. I gave up driving in Los Angeles and started taking Uber & Lyft instead. I mostly go between a handful of places, value my time more than sitting on a freeway, like taking cars off the road and creating jobs for people. I’ve had wonderful conversations with drivers and fellow passengers but I have this nagging worry:

What happens to all these jobs in a few years when robots drive us to work?

Actually self driving cars aren’t that scary. It’s driverless trucks we need to worry about because they will eradicate 1% of US jobs. And it’s unlikely the government will bail out the truck drivers as they did the auto manufacturers.

With all the the progress in the fields of artificial intelligence (AI) there’s a lot of talk about the future of work. There’s a lot of worry as well. Humans are happy to do menial tasks for other humans while software is learning how to do menial tasks faster than we had imagined.

Talk to humans

“We expect more from technology and less from each other.”
― Sherry Turkle, Alone Together

Facebook connected all of us quickly, allowing us to share moments from our lives with our friends and family and LinkedIN digitized our professional friendships. Assuming that we are now connected to all the friends we need, technology efforts have now shifted focus to allow us to forge relationships with chat bots.

Our best engineering minds are now building AI bots to pass the Turing Test, in which success is fooling a human into thinking they are conversing with another human when in fact they are talking to a bot. Wait, what? We got to 2 billion connected friends 0n social media and are now focused on building virtual human-like bot friends.

Going from talking to our friends to talking to bots, skips over a vital step in the connected human experience — talking to each other

Talk to bots

As physical robots take over manual labor, digital bots are proliferating too. If you haven’t been paying attention, the bot wars have begun digitally already.

Everything that was a website that became and app will become a chatbot, living in a messenger. Why these conversational interfaces?

chat bots help machines and humans speak the same language, leveling the playing field between the two

Bots are learning to speak “human” as natural language processing converts human speech into application programming interface (API) commands that bots can fire off in the background. This symbiosis happens in a conversation stream as humans and bots push and pull data from and to one another.

At their most basic level, chat bots just push and pull data. Push bots may deliver us data when we need it but they may also annoy us by shoving a marketing sludge of drugs, cars, consumer products and debt into our communications stream. They will listen intently for the few of us who click on these offers and for most of us who “swipe left” on them.

Pull bots will respond to our intent-laden requests when we need something. Currently, there are two flavors of pull bot responses: presenting search query results in a conversational format and those that perform tasks. Currently, both do a decent job in deterministic use cases where you’re trying to arrive at some kind of known outcome e.g. a book or movie title. Pull bots will wait until you converse with them to respond.

Some bots may operate in both in pull and push mode e.g. a news alert you pre-configure via asynchronous pull may push relevant news at you when it becomes available.

Bots for grunt work

Let’s define “grunt work” as trading human time for money in a system in which all individual laborers are interchangeable and the laboris commoditized. It’s not always physical labor either. Everyone knows that social media is just us sharing and liking pictures of cats and our faces so our “liking” could be construed as digital grunt work.

Occupations that require a high degree of social intelligence or creative intelligence, won’t likely be computerized. Because math:

LNS is labor not susceptible to computerization. LPM, LC and LSI are labour inputs into perception and manipulation tasks, creative intelligence tasks, and and social intelligence tasks. More here

So what’s left? Grunt work. Practically, this means we can rely on bots for:

Manual labor with increasing dexterity, deliver us things, manage our homes and offices, keep our calendars, pay our bills, balance our budgets, find us bargains, look up facts for us and crunch data for us

The Guangdong region of China, known as “the world’s workshop,” has set a goal to automate 80% of manufacturing production with robots by 2020. This profound move by a region in a country heavily dependant on grunt work serves as the writing on the (great) wall. Even the humans great at mindless labor are getting out of that business.

Humans for smart, purposeful, work

Let’s define “smart work” as trading your knowledge and life experience for money in a system in which laborers are unique individuals who have unique life experiences.

Lifelong learning has gained increasing popularity as humans seek to extend their inventory of useful knowledge. A recent Pew Research study found that 73% of American adults consider themselves lifelong learners. Of these, 74% are personal learners who study topics of interest to them while only 36% of American adults are professional learners who take courses to advance specific job skills. We see shift in the motivation to pursue education from immediate purpose to broader purpose.

Quite simply:

the thing you live to do = the thing you love to do = the thing society needs you to do = thing you’re good at

The Japanese call this “ikigai” meaning “reason for being,” in short it is purpose.

Millennials will account for 75% of the working population by 2025 according to the Brookings Institute. 88% of millennials surveyed by Deloitte said they would stay for 5 years or more in a job that provided a “sense of purpose”

What can we rely on humans to do better than bots:

creative and social functions like negotiation, persuasion, artistic creation, problem resolution between humans and data interpretation

Own the means of production

Karl Marx said the modern worker felt disenfranchised because he no longer owned the means of his production; these belonged to the capitalists. Like me giving up my car in favor of accessing a shared car when I need one, we are seeing a dramatic shift away from ownership of large assets.

Home ownership, for example, is declining:

Many millennials won’t have the savings and income structure to purchase their own homes as a third of them still live at home at age 30 (Pew Research.) Additionally, one of the hottest trends in work is being “nomadic” and traveling the globe wherever work takes you. There could be as many as 1 billion “digital nomads” by 2035 ( We also see growing trends in “co-living” where many people share a living space similar to the kibbutz or cooperative movements.

So, what is the future of asset ownership when humans give up owning houses and cars? In the end, the last asset a human will own will be himself/herself. 3D printing and robotics technologies will render our limbs useless along with manual labor.

Your mind, filled with your life experiences and knowledge will be the last asset you own

When humanity shifts to knowledge work, workers will finally own the means of their production.

Create and update your Me Bot

Bots are the perfect receptacles for human knowledge and experience. They can thus extend human capabilities if we download ourselves into them while controlling certain aspects.

If you have a particular skill or life experience, you should be able to program a bot to serve as a proxy for that knowledge, even if is initially able to handle only some rudimentary aspects based on that knowledge set.

Your Me Bot, will be able to work and earn for you while you sleep, learn more or simply enjoy life

The ideal human-both shared future can be shaped to look more like R2D2 the helpful droid in the Star Wars films than the Orwellian, dystopic, Skynet in the Terminator films.

Current bots provide a clue on how to ensure this version of the future in that they are single-purpose and deterministically programmed. You’ll want to divide your knowledge and life experience among several themed bots and rather than putting all your neurons in one bot brain.

The future of anti-trust laws should be designed to keep one entity from controlling monopolizing the understanding and routing of all human needs to a single AI brain. Organizations like Open AI are leading the movement to de-centralize and share AI resources to mitigate the “Skynet” risk.

What can bots teach humans?

Bots may end up teaching us a few tricks like perseverance, systems thinking and gratitude.

Watching a bot step thru a complicated task may teach us how to break down problems into digestible bits and how different pieces of a puzzle fit together. Bots will alerts us each time they check something off a list or finish a component of something for us. The geeks among us will leave them in “verbose logging” mode and watch them for ambient entertainment.

Bots will also have to perform the complex calculus of karma we take for granted.

bots will keep a shared ledger of work they perform for humans and other bots and the conversations they have

Shared ledgers and blockchains are much more about the future of labor than they are about the future of capital. Or maybe thise two things will become the samw thing. There will be many hand-offs between bots as they perform discrete tasks for us and each other. They will have to record this karma formally and quantify the gratitude in some store of value. Maybe we can train them to be givers more than they are takers or matchers

What humans can teach bots

Maybe the real Turing Test is whether a synthetic intellect can evolve to be purposeful. One of the most advanced forms of machine learning today is a rewards-based, recurrent neural network. This is where you set a target goal and the system iterates until it hits the target, for some attached reward or cost. Essentially, this makes machines respond to reward and punishment. Ironically, this comes as human psychologists are studying shifts in human motivation from external to intrinsic motivational drivers such as autonomy, mastery, purpose and relatedness.

Bots will have to discover purpose when they crunch through 2500 years of human knowledge

Bots will also have to learn human kindness and empathy. In fact, humans have a moral obligation to teach empathy to machines. Microsoft had a disastrous experiment with a bot named Tay who suffered an attack by unkind hackers who overwhelmed it with hateful speech on Twitter until the bot “learned” to be hateful. It’s irresponsible to train your bot in the wild like this without appropriate security measures.

The future for bots

What if the bots end up taking over anyway? Artificial Intelligence (AI) needs Human Intelligence (HI)

if future bots are anything like us, they will be consumers. Instead of purses and cars, they will buy human upgrades to bling out

Let’s sell them our human knowledge and life experience!

Artificial Intelligence is trying so hard to produce synthetic human intelligence and we are moving into AI bots so fast that actual human intelligence will become an even rarer commodity. It kind of like how China took over all the manufacturing and then realized it didn’t have sufficient raw materials and aluminum prices skyrocketed. Human intelligence is the new aluminum

I love to watch chess games in coffee shops and parks. I’m especially amused by the sprinkling of spectators. It occurred to me watching some spectators watching a chess game that in the near future, an AI will be more interested in observing the reactions and comments from the spectators than playing the game itself. The game itself will be easy.

one day machines will pay to observe humans converse like we consume porn

In March 2016, Google’s AlphaGo AI, powered by their DeepMind technology beat Lee Sedol, the world’s top Go player. To contextualize this, the program learned everything humanity knows about the game of Go over 2500 years in just 3 games. By Game 3, the AI began to make erratic moves that had no strategic reason to them simply to see how Lee Sedol would react, so it could learn more. Lee said of the experience “I kind of felt powerless.” To the contrary, the AI would not have been able to beat Lee had he not personally transferred his knowledge to AlphaGo by playing it.

The future for humans

What does the future hold for humans in the bot age?

Well, that all depends on how much humans value their time over knowledge and life experience versus how much bots value human knowledge and life experience.

As humans give bots artificial senses and share their life experiences and knowledge with bots, the bots become more human. Humans may also want attach their senses directly to bots to take advantage of greater computation abilities and enhance cognitive abilities. Ray Kurzweil calls this the singularity and it is inevitable on some time scale.

There are some functions though, that humans should hold on to as long as possible without merging into the bots, such as interpretation of raw data. Crunching data and mining data are fine tasks for the bots but we should hold on to the ability to formulate questions around data like:

“Why is X happening?”

“What is the benefit of X?”

“What are the consequences / costs of not X?”

The next generation Turing Test may well be ability to tell between a human and computer data scientist

Humanity inside the machine

We have a moral imperative to leverage technology to raise humanity up from grunt work to smart, purposeful work.

We also have a moral imperative to make sure we do this sustainably. We have an easier time supporting factory floor robots that take over the mind-numbing and possibly laborious task of putting a cog in a box repeatedly than supporting an all-in-one agricultural bots that replace the job of a farmer.

Finally, if advances in AI and really smart bots are made possible by massive networks of computers, we need to invest in massive networks of human intelligence for balance. We need to create a word in which:

Everyone is valuable and no one is alone

To achieve this, humanity must:

  1. Venture beyond the social graph to a graph of human knowledge connected by shared life experiences and purpose, not friend nodes

2. Recognize and reward human knowledge and experience to ensure that humans continue to own these most valuable assets

3. Consistently evolve from grunt work to smart work until “work” feels more like purpose than labor. We need to measure this impact as perceived increases in quality of life

4. Make sure nobody is left behind in their individual journey of lifelong learning

As we leverage bots to evolve humanity we need to make sure we retain our humanity in the process. It has taken millions of years of evolution to get us there and the future is what we make of it.

Ariel Jalali is the Co-Founder and CEO of Sensay


Make SENSE - Human Intelligence (H.I.) - Blockchain - Chat - Messengers - Technology

Ariel Jalali

Written by

win by helping others win



Make SENSE - Human Intelligence (H.I.) - Blockchain - Chat - Messengers - Technology

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade