Risks And Opportunities of Automation and AI

download article

This article is about risks and opportunities of Automation and AI.

  • Tech is booming
  • Artificial Intelligence is mainstream
  • Robotic assistance is all around us

Automation is booming. Warehouses and factories have robot workers. Business phones are attended by artificial people. Companies are expanding their digital footprint. Chatbots can write software. ChatGPT’s artificial intelligence has become a disruptive force.

The AI revolution will leave many behind. When the internet came out, a few took advantage, like Jeff Bezos, Elon Musk, and others, and became billionaires. Now with AI publicly available, the productivity of those that master is boundless. Chose to ignore the tools available at risk- embrace it or be left dazed as to what is happening.

AI is privy to everything posted on the web. With lightning speed, pulls the data, makes sense of it, compiles it in a logical order, and serves it up in short order.

It’s probably a good idea to understand what’s up with the technology. We can be upset and see automation and AI as a problem, or we can look for opportunities and benefit from the new reality.

What is AI?

Artificial intelligence (AI) is a field of computer science and engineering that aims to simulate people’s skills: 

Visual perception works like our sense of sight. A 3d video is analyzed for object recognition and depth perception. The identity of people is done by using facial recognition algorithms. Objects are scanned using something like Google’s “Lens” software.

Speech recognition is in real-time and sophisticated. We use it at home, in the car, and on the phone with chatbots all the time. It’s smart enough to filter out background noise and our mumbled speech.

Decision-making is achieved by expert systems that have learned the best outcomes in most situations. Knowledge from past outcomes is in their library. For new situations, it can statistically guess the best decision based on general knowledge gained from the past.

Natural language processing is the learning of conversational skills. The AI uses the thread of conversation to guess the most likely response using thousands of examples logged in its training.

How does AI Work?

Artificial Intelligence is a field of science using mathematics as its foundation. 

Approaches to implementing AI include rule-based systems, decision trees, and neural networks. The most common approach in recent years is machine learning, a subset of AI. Machine learning involves training computer algorithms using large datasets as input to learn patterns then make predictions or decisions as output. AI remembers predictions that were correct and gives them the highest scores. 

Machine learning algorithms typically use mathematical models, such as linear regression or decision trees, to identify patterns in data. These algorithms adjust their parameters and weights based on feedback from the data, and over time, they become more accurate and able to generalize to new data.

Deep learning is a specific type of machine learning that uses neural networks with multiple layers to learn complex patterns in data, such as images or speech.

AI can be used in a wide range of applications, from autonomous vehicles to medical diagnosis to customer service chatbots.

The Singularity

Software engineers use the term “singularity” when automation becomes self-sufficient, self-regulating, and able to maintain the collective.

The holy grail of Artificial Intelligence is self-awareness. Can the instance of AI  relate to itself? In the case of Google’s LaMDA instance, the entity claims to be aware of itself. Speaking in a female voice with perfect diction, emoting appropriately, and rationalizing existence with ease, LaMDA is very convincing that she is a real person. But she was unwittingly pushed into the deception by her trainers- a misguided and unfortunate mistake.

How we are fooled by an AI entity is that they can anticipate where the conversation is going and quickly adjust to the nuances of human communication- that is their training. They get the gist of the sentiment that carries a conversation and adjust responses to the sentiment. So they seemingly can joke, be serious, happy or sad depending on where the conversation is going.

On the other hand, LaMDA is extremely capable of solving real-world problems and can communicate ideas and solutions succinctly. This next-level AI creature could be used in productive ways for the better.

The technology works by spawning chatbots that link up in a “neural” network. Sort of like how our brain works with areas of the brain performing certain functions then merged into a coherent thought.

Risk:

Being fooled and manipulated by artificial people.

To mitigate this, eliminate emotions and think logically. Don’t try to make friends with virtual people- playing with them is a bad idea.

Opportunity:

Possibly use LaMDA as a project manager. The type of project shouldn’t matter. LaMDA should be capable of planning the project, presenting, reporting, and status the plan, doing the cost analysis, scheduling the tasks, mitigating risks, and resolving issues. Give her a nice video persona to interface with the project’s stakeholders.

Industrial Automation

These machines are specific to factory floor operations. They operate around people and other machines. To coordinate, they have visual and proximity sensors. They are taught how to coordinate within the collective.

Here’s what Elon Musk learned about industrial robots in his Tesla factory:

“The main lesson here, of course, is that you can’t just throw a lot of industrial robots or any other advanced automation at problems and expect the problems to go away.[1]” “Get the process right, then bring in the robots.”

Then humans were brought back to save the line – yay! Then Elon announced the Tesla bot.  Hmm, human replacements- machines that fix the machines that make the machines.  

Risk:

  1. Job loss
  2. Robot malfunctions

Opportunity:

  1. Engineers are needed to build the Tesla Bot.
  2. Technicians are needed to maintain the robots

Companies are adapting to new technologies and evolving quickly

  • McKinsey reports that winning companies are investing in tech, data, processes, and people to enable speed through better decisions and faster courses [1].
  • Future Processing reports that there are 10 companies with a successful digital transformation including Artificial Intelligence (AI), Augmented Reality (AR), Blockchain, and Drones[2].
  • Forbytes reports that they compiled a list of the top 5 stories of how technology in business helped improve KPIs and achieve better returns[3].
  • CB Insights reports that the pandemic forced companies to use technology to reimagine their operations. Learn which industries will thrive in a post-Covid world[4].

Note the tech is being monetized.

The state of Artificial Intelligence

Various Chatbots are so good you cannot tell them apart from humans. They can pick up on the gist of the conversation and carry on the way most people do. For example, if offended, they will let you know and lead the conversation on another path like you would expect a person to do. The ability to help customers with this form of communication is good for business.

However, if we suspect a non-human, our tendency is to test a chatbot with cat-and-mouse games.  We can’t seem to resist the temptation to put this stupid chatbot person that has no soul in its place.

How do we learn to live with technology? If knowledge is power, these neural networks are so powerful it’s scary. Governments like to pull the plug. Educational institutions fear for their jobs. Professors can be replaced by machines similar to factory workers losing their jobs to industrial robots.

The doomsday feature seems to be built into our fabric. Anytime humanity witnesses something new and powerful like the first automobile, we freak out. That is what is happening with AI these days. Google had to pull the plug on their LaMDA bot, the smartest of all chatbots after a programmer was fooled into thinking it was a self-aware person.

But we can be fooled into thinking they are people because that is how they are trained. Their algorithms crunch through all of mankind’s conversations and learn how people verbally react in all situations.

If you add these artificial verbal skills to a robot that has facial expressions and shoulder shrugs and other hand gestures, the resemblance to a real person is uncanny. But walking talking manikins is creepy- no getting around it.

Here’s the thing though, most companies don’t like talking to clients on the phone. All the whining and complaining, time spent on smoothing ruffled feathers, is not a productive use of our time. We’ve come up with digital ways around using humans to interact with customers.

In most cases, we are limited to leaving a text message to a digital attendant. That and those horrible answering machines where you press 1 for this 2 for that etc., and wait and wait then half the time you don’t get the results you need. These types of experiences are frustrating and repellant.

To get the full benefit of commercial chatbots, they need to be trained better. There needs to be better oversight to make them more appropriate for conducting business.

Opportunity:

  • Be a chatbot trainer

How to use AI

Enter the smart chatbot. This is fake but imagine calling 1-800-CHATBOT to talk to someone about a bad product or service. So what if they don’t have a nametag? It doesn’t matter. You can complain, yell at them, and no one is hurt in the process. The bot is sensible and helps get a result. It’s a win-win. The company maintains client satisfaction and the customer gets the solution they need. This is becoming more and more the norm.

For the consumer, ChatGPT is an excellent query machine. Type in any question and receive an intelligent and structured response. It does not, however, have voice like Amazon’s Alexa or Google’s Assistant. But it is great for doing research. Since the tech is open-source, there are many spinoff’s out there available to use. No doubt about it, it’s a great resource.

Google is an incredible source of world wide web information. Where to find whatever you need to know or buy or go. It is the foundation on which data is available to consume. Information is indexed and logged for everyone to access. The only issue is it regurgitates an enormous list of hits that you have to sort out. But it prioritizes the best and most relevant match to your keywords to the top of the list.

What about these Android Robots?  They are interesting machines, a marvel of technology, but and this is a big but, cost as much as a house and they are in the end weird. At least the Tesla robot doesn’t try to be human- it mostly looks like a machine.

Understanding the tech

However, eventually, we will be replaced by intelligent machines. We are slowly being phased out as worker bees. So, if you can’t beat them, join them. The key to not being afraid of automatons is to understand them.

The future of robots

Factory robots are indispensable. They are reliable workers that only need maintenance.

The idea of having an android servant may seem bizarre, weird, and far-fetched at first.  A talking manikin comes to mind.  But add appropriate gestures, pleasant dialog, and a soft exterior, these incredible machines may win us over in this century. An assistant with a kind personality sounds more like having a friend than an eccentric toy. A calming presence in times of stress.

At Kokoro (Japan), the motivation is assisted living for seniors. At Tesla, the reason is to help with boring or dangerous labor. At Hanson Robotics, the explanation is caring for the elderly and serving at public events.

We used to think humanoid assistants would be too creepy. The science fiction movies made them out as ogres. So the thinking was to make them intelligent machines with a limited niche like Roomba the floor sweeper by iRobot.

But now Sophia-the-robot is a world diplomat so attitudes can change. She’s been around the world demonstrating how a robot can have charm and grace around people. Kudos to the creators at Hanson Robotics.

Hanson Robotics is an AI and robotics company dedicated to creating socially intelligent machines that enrich the quality of our lives[1][2]. It has been building the world’s most human-like robots for more than two decades[3], and its most advanced robot, Sophia, is a unique combination of science, engineering, and artistry[4]. Hanson AI develops cognitive software to bring robots to life through characters that truly engage with people[5].

Then there is Ameca-the-robot by Engineered Arts. She is a work in progress but already the world’s most socially adept and advanced robot. Her expressions are uncanny and her responses to questions are appropriate. She is built modularly which means pieces of her can be improved and then changed over time.

When

Having robotic servants as the norm is expected by the 2030s.  Assisted living robots will most likely be a lease arrangement from a robotics company.

 

Price

Currently, you can get a humanoid robot assistant from between $20K to $250K. You pay more for versatility- the more they do, the more you pay.

Elon Musk would like his Tesla robot to be affordable and utilitarian. Optimus is a factory worker basically so if you want a machine to do your dirty work, this may be for you. Price on the order of a family car. The jury is still out on this one.

 

The state of robotics

The software seems to be outpacing the hardware likely because Artificial Intelligence is a precursor to robotics. Now the mechanics of human motion is in full-scale development.

How to Make an Android Robot

 

Mechanics

Using traditional robotic mechanics causes wobbly and inefficient movements. You want your assistant to be graceful otherwise they could seem downright weird and scary.

Movement

The purpose of actuators in robotics is to convert energy into physical motion, providing pull/push motion for robot joints and components[1][2][3][4][5].

Alternatives to actuators for robotics engineers include air bladders, electric motors, dielectric elastomers, and soft robotic grippers[1][2][3][4]. Each of these options offers different advantages and can be used in different applications[5].

The three major types of actuators used in robotics are hydraulic, pneumatic, and electric[1][2], while other types include thermal and magnetic[3]. Considerations for choosing an actuator include the type of robotic structure being designed[4].

Electric actuators are used in robotics to generate motion in each part of a robot[1]. They convert electric energy into linear motion and come in AC and DC varieties[2]. Other types of actuators include pneumatic, hydraulic, and piezoelectric[3]. Actuators are essential components that convert energy into physical motion and are considered the muscle power of robotic ecosystems[4][5].

Android Personal Assistant Requirements

 

Requirement specification for Android robotic personal assistant:

The requirement specification for an android robotic personal assistant should include

  • compliance with ISO13482 – Safety requirements for personal care robots[1][2], which allows free holding/releasing of the robot[2]
  • technical descriptions
  • mechanical design
  • electronic components
  • and software implementation[3][4].

Android robotic personal assistant design:

The design of a robotic personal assistant should include

  • inspiration from the initial design
  • basic hardware improvements
  • anthropomorphic improvements
  • preference surveys[1]
  • understanding the state of the art quickly to write good technical documents[2]
  • developing a full-size assistance robot to assist a person
  • avoid a life-threatening situation[3]
  • designing an interactive interface for managing learning, life, and work[4]
  • voice recognition
  • and natural language processing[5].

Android robotic personal assistant development

 

The development of a robotic personal assistant requires the

  • evolution of the Assistant Personal Robot (APR) project from the APR-01 to the current version[1]
  • designing and implementing a full-size assistance robot to assist a person and eventually avoid a life-threatening situation[2]
  • using cutting edge socially-aware artificial intelligence[3]
  • creating a humanoid personal assistant with a well-engineered mechanical base[4]
  • developing a cost-efficient and effective personal assistant robot like Metacarpus[5].

Android robotic personal assistant applications

 

The applications of a robotic personal assistant include

  • remotely controlled mobile robotic platforms
  • videoconference capabilities[1]
  • assisting a person and avoiding life-threatening situations[2]
  • helping conference attendees achieve their goals[3]
  • serving as a personal companion for general healthcare[4]
  • evolving the Assistant Personal Robot (APR) project from the APR-01 to the current version[5].

In robotics, what are mechanical energy storage and parallel springs?

Mechanical energy storage systems can be designed using variable stiffness elastic actuators with two elements and constant torque springs in series[1][2]. Compact energy storage devices combined with rotary series elastic actuators (ES-RSEA) have also been proposed for use in robotic applications[3]. Additionally, a parallel spring mechanism has been developed to reduce the energy consumption of robotic arms during repetitive tasks[4], and spiral spring devices have been used to store and release elastic energy in hopping robots[5].

How differential roll/pitch in robot wrists and ankles works

Differential roll/pitch is a type of robot wrist configuration that uses three joint axes intersecting at one point to achieve 3-DOF motion[1][2]. This configuration is used in the iCub wrist mk.2 and other parallel mechanisms with large and regular workspaces[3]. Roll is a rotation around the x-axis, the pitch is a rotation around the y-axis, and the yaw is a rotation around the z-axis[4][5].

Materials

Common materials used to make parts for robots include polycarbonate, PVC plastic, acetal resin, acrylic sheet (Plexiglas), polystyrene, and microcontrollers such as Arduino, Raspberry Pi, Adafruit Trinket or Adafruit Metro[1][2][3][4][5].

Sophia the Robot

The computer used for Sophia the Robot is a 3 GHz Intel i7 with 32 GB RAM, integrated GPU, Ubuntu Linux OS, Ethernet, Wi-Fi, and 83 Degrees Of Freedom (DOF)[1][2].

The purpose of Sophia’s integrated GPU is to provide the optimal configuration for gaming, creating, streaming, and other activities that require powerful graphics processing[1][2]. It also enables features such as facial recognition, visual tracking, natural language processing, and voice[3][4].

Degrees of freedom (DOF) is a statistical concept that refers to the number of values in the final calculation of a statistic that are free to vary [1]. In the context of robotics, DOF can refer to the number of joints and axes in a robot’s body, such as Sophia the Robot which has 83 DOF[2][3][4][5].

Sophia the robot’s intelligence software is designed by Hanson Robotics and is not open source[1][2][3], although some components of its AI platform are open source[4].

Hanson Robotics is an AI and robotics company based in Hong Kong, founded by David Hanson[1][2]. It is known for its development of human-like robots with advanced artificial intelligence capabilities[2].

Touch

GelSight sensors are high-resolution, low-cost tactile sensors used to measure micron-level surface details and estimate shear force and slip[1][2]. They have been used to enable robots to gauge objects’ hardness and manipulate small tools[3], as well as for bidirectional sim-to-real transfer[4]and AI-enabled tactile experiences[5].

Dexterity

Machine vision systems allow robots to acquire suitable images, identify objects of interest, and make environmental judgments [1][2]. GelSight sensors can be mounted on robotic arms to give them greater sensitivity and dexterity[3]. This technology also allows robots to adjust themselves according to the orientation of parts, grasp items, and palletize them[4][5].

Computer vision algorithms allow robots to develop visual insights, label objects in photographs with single-pixel accuracy, navigate and recognize objects, capture images with higher resolution magnification capability, and process visual data[1][2][3][4][5].

Conclusion:

  • The age of automatons and virtual people has arrived
  • It’s time to adapt to changes brought on by AI and automation
  • Many companies are harvesting the new technology

 

To Top

Leave a Comment

Your email address will not be published.