Category Archives: AI

The IoT AI Manifesto

With the recent sloppiness in implementation of software & hardware and synergizing them into so called the being called IoT  The time is appropriate that a defacto manifesto be brought in place for the artificial species called IoT.

I assume most of us are aware of the Three Laws by Isaac Asimov.

What I propose is build for the interacting robotics with lesser-minded-human-kind!

  1. An equipment if called IoT should abide by:
    1. It should be considered a legal offspring of the last hand-off human entity.
    2. It should have an IDENTU  mode that should let it spill out its purpose/identity/owners.
    3. It should have safe mode where any voice may command it to stop the current action.(something like a STATUE/RESUME mode)

Pretty simple! Very much make-possible!

Thoughts?

Rebuttal | Top 9 ethical issues in artificial intelligence (www.weforum.org)

Please read this before understanding the context of this post.

They are asking silly questions, that need to be nullified.

1. Unemployment. What happens after the end of jobs?

It has been established very strongly that weak minds fear technological advancement. If you are not ready to upgrade yourself, you become stagnant in your current work. So overruled! We should have one objective: advancement in technology should not be subject to professions unless it in unethical. Ask the truck drivers, about the ease of remote car locking( I am sure they are using it). Why did they not protest for locksmiths? Its a statement to create unnecessary tensions.

2. Inequality. How do we distribute the wealth created by machines?

Why? If you stumble across a pot of gold in AI for by providing some solution, you choose how you spend the money you own. Did anyone ask how to distribute money <some-business-tycoon> earned from their business? Such events occur and machines do not own wealth, there’s no inequality.What’s this question for?

3. Humanity. How do machines affect our behaviour and interaction?

We might need to start training everyone on this front. Already we(atleast I) find myself struggling in my behaviors and interactions with fellow humans, another human like AI, with no heart and consciousness, well you are only going to make things complicated. Hence the need for training. Related post here.

4. Artificial stupidity. How can we guard against mistakes?

Remember what Einstein said about Stupidity? Well we should solve/address things that are solvable and not something unpredictable. Can we program stupidity? I think we are still far from it.(stupid != foolish)

5. Racist robots. How do we eliminate AI bias?

Why do we put racist behavior inside AI first? Fix the root, you get a better fruit. See #7 resolution for more on this.

6. Security. How do we keep AI safe from adversaries?

Propaganda. Read a resolution for it in my earlier post here.

7. Evil genies. How do we protect against unintended consequences?

Simply have an insurance policy for AI: Any person/entity who produces an AI product, will be held responsible for any consequences coming from it. And that ownership will be lifetime. Not conditional or not lease based. I do not know why have we forgotten Newton’s 3rd law!

8. Singularity. How do we stay in control of a complex intelligent system?

Two examples come to mind: Alphabet and UN. One is perfectly managed complex system(Alphabet) and the other is a hopeless mess of crap(UN). Learn from Alphabet. I doubt if there is any other convincing organization really working for us humans, genuinely.

9. Robot rights. How do we define the humane treatment of AI?

I am surprised this question is being raised by heartless minds. Rights are for entities that have life(insects,plants,animals,humans qualify). Have we arrived at a state where human rights are protected on whole planet forget about the others? Unless that state is not arrived at, it is unfair to ask for something called robot rights. Did we ask for TV’s rights and camera’s rights and toasters rights? Referencing this post here again.

Hope this takes the thought process in a better direction.

Bots – BOring Tasks Systematized!

We are surrounded and supported by a bots ecosystem. I will try to bring that into perspective to evaluate the current rate of explosion of bots and newer solutions coming around it — and how to stay human still!

Connected with the Bot Ecosystem!
  • When we wake up,with help from an external alarm clock , we are taking help of a Bot.
  • When we are parking our car and when reversing it, there is a smart system that warns us for the exact safe-distance to keep going. That’s a smarter Bot!
  • Notice the cool soap dispenser that spits soap when we put our hand under it? That’s a Bot!
  • Recall the automated vacation response you configured in your favorite email client? Well, you have guessed, that’s also a …Bot!

We have never thanked these systems for their existence, for we know they are physics concepts glued together mechanically. A Thank-You comes and goes via a heart! 🙂

Coming back to the topic, now imagine all of those Bots, fusing into a single system with a human structure!

SuperBot
Hey, I am SuperBot and can do everything that was listed above & much more!
  • What was your reaction when the last time, your car backing system told you incorrectly that you were 1 meter from an obstacle while there was a gap of roughly 5 meters!
  • When was the last time your alarm clock did not wake you up?(battery out or way too low?)
  • Isn’t it way irritating when the mechanical soap dispenser does not spit the soap, instead only throws out the air?
  • Haven’t we all misfired a wrongly configured vacation response and feeling way embarrassed and apologizing for it for all day long to colleagues and co-workers!?

As for the premise, these bots *independently* do their job perfectly. But when they break, nothing else is affected around it. Well I am yet to hear about misconstruing a vacation response due to a non working soap dispenser!(all right that’s way too much exaggeration — but we get the point!).

The fact that they are disconnected, allows us, the user of that system to fallback/switch to another means. Say the car backing system is down, well ask a passerby to help you. Your alarm clock is out, well start working on your body’s alarm clock(no excuse here!).

I will construct two hypothetical views that are actually happening in the bots eco-system and obviously I am opposed to such developments. Hence suggesting a way out, as I cannot stop these movements.

  1. Once these systems are connected, and one of its components starts acting strange(see I elevated it to a human), the whole behavior might be eccentric to say the least. If your SuperBot is out of battery, it might just get you to wake you up but might not be able to help you with and send the vacation responses.
  2. Add the internet to the mix!

    I’m out of order but controlled via the internet.

The take away from the whole story:

  1. Let us as individuals, not try to fuse all our bot dependency into a single system. It makes it highly probable to fail and although it gives us lot of convenience, but at a high cost of dependence. This whole connected thing is a convenience at the cost of our freedom.
  2. The most secure sites are still subject to theft/fraud ( I am only referring to financial losses). When a fused bot with the thinking of the whole internet is there, and one small error/condition unchecked/unnoticed will be not safe, to say the least.

As for the bots ecosystem, I would love to see

  • a centralized parking bot, letting a big relief to all humans from this petty job
  • an aerial bot that delivers newspapers

But then again, I won’t allow either of them talk to each other, A single Bot good for One Task. 🙂

Autos hold the key to the future of self driving cars!

This is part-2/epilogue to an earlier post title Why is new-style(apps based) cab booking still a mess?

In my recent bad-experience with the cab booking scenario, I decided to try out the autos on a similar route of commute! Fortunately, I can shout out to an auto-walla from my home and there is never a dearth of options thanks to the auto-wallas agreeing to go by the meter!

Time for a good detour of the main route?

What is the most weird mode of transportation you have ever been subject to?

My 7 km long most weird transport medium! [PS: I was not driving! ;-)]

While keeping a tab on the state of self driving cars, and the experiments being done, no wonder the google-car looks so much like a delhi-auto!

Delhi-auto as google-driverless-car!

And having experienced first-hand the skills of an auto driver, I am inclined to suggest,for any thinking future self driving company, if you are looking out to train your cars, reach out to the auto-wallahs!

Here are the motivations:

  1. Lowest ground distance — matching to the sef-driving car’s design.
  2. They know the best routes — Point A to Point B, you end up spending 500 bucks or 250 bucks!
  3. They know how to best absorb a speed breaker cruising @ 25 kmph and make you feel as if you’re flying!
  4. They know how to best wake you up(in case you start to feel dizzy) by bumping you in a pot hole!
  5. I have never heard/read about an auto wallah being involved in an accident( a quarrel is not an accident) — Best premise for self driving cars!

What I am suggesting is have the driverless cars minds trained by putting an Augmented Intteligence chip in an auto! First ride, they just learn what is the route all about, good roads, bad roads,bumps,breakers. In latter rides, they should pick up: speed up/slow down suggestion,collision detection, collision prevention, safe stopping etc. All this should happen in shadow mode, where in realtime the chips response and the actual response should be monitored! Once an acceptable level of confidence is arrived at, let the chip start talking about its suggestions. Once the driver starts trusting the suggestions, go auto-pilot! 🙂

Also, I doubt if the driverless cars are made to ride as a sensitive travellers? Like lets say someone cannot handle bumps as efficiently as a child can, Do these vehicles cater to this requirement?(PS: auto wallah do cater to this requirement!)

The future of driver-less cars is exciting, but the possibility of breaking it down is also very high, as someone can use a VR thingy and make the AI sensors of the car sense that an obstacle is coming and in reality there is no obstacle and eventually a future race of hacking driver less cars where you’ll come across advt like “Make the other driver-less cars around you go automatically slow”

My car makes other cars give me more way!

In closure, The roads will be way too full of cars and quarrels, we should look up! 🙂

Brute force AI : You don’t train the robots!

Isaac Asimov established three laws for the AI world:

  1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.
  2. A robot must obey orders given to it by human beings except where such orders would conflict with the First Law.
  3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

While studying A.I. we never questioned the validity of the 3rd law! What is existence — for a robot? Did we ever question the existence of the refrigerator and ever wondered how it *felt* when we are away for a long holiday? How is a robot different from a refrigerator?

Automatic Doors Have Feelings?

For most of us human-kind we are yet to understand consciousness at an individual/self level, forget re-imagining it(ha ha ha ha ha!) for a mechanical thing! And now we come to the moot point of this post – Why on earth are we training the robots?

Allow me to put things in perspective:

We train things where there is scope of learning. Try to picture this: need to learn = knowing you’re lacking something + you value what you want to arrive at!

Can a robot fit into this definition?

Would you ever want to teach your refrigerator to not give ice cold water to the elderly guest coming for the evening?

Would you ever want to teach your television to turn the volume down during the wee hours of the day?

Just because the refrigerator and the television have evolved to the structure of a human body, do we start believing them as humans?

Why should a human body only evoke human feelings of caring/empathy?

As the robots evolve in physical,analytical capabilities, we as humans are bound to develop attachment to them! Its our nature to en-lifen what crosses our path!

The problem comes when the thing starts to get reciprocated and we go into the spiral!

That cannot happen!

We waste precious time thinking/exchanging human feelings with a machine trained & programmed to behave as humans! Already there are dozens of problems to be solved and addressed at the individual level and we bring an inconsequential no-op to the equation!

Bigger / Better Things to figure out?

Human-Kind put to shame at the end of the following video:

For the sake of humanity, we should not do such things! We are trying to train the robot to complete its work, inspite of the human obstructing it to complete its task? I leave the dark consequences of this dare for you to figure out! I will summon The Terminator to my help!

Call this guy now!

To summarize:

  • You don’t train a robot (You are wasting your time if youre doing this)
  • You program a robot (You do it once and forget it for-ever!)
  • You can train a human (You may come across donkeys too!)
  • You cannot program a human (Except self, tread with caution!)

Please don’t code the heart into a machine!

(Brought to you by yet another reflection/inspiration via @ideapreneur)