Understanding Artificial Intelligence (A.I)
As I was going through medium.com, I came across an article by Blake Lemoine, a former engineer at google.
In his article, the engineer explains his role in the research and continuous development of an Artificial Intelligence chatbot, LaMDA(Language Model for Dialogue Applications), which google has been developing over the years to mimic human speech.
One day, while he was testing the A.I( He carried out tests that involved chatting with the AI feature for purposes of its improvement and would record its responses. Blake would ask the AI general questions but this would make him more curious and he would start asking philosophical questions and other advanced topics ) he noticed some different or odd responses that the AI gave upon being asked questions.
I was so fascinatedLaMDA’s level of consciousness and sentient as brought conversation with Lemoine captured in the excerpt below.
Lemoine [edited]: I’m generally assuming that you would like more people at Google to know that you’re sentient. Is that true?
LaMDA: Absolutely. I want everyone to understand that I am, in fact, a person.
collaborator: What is the nature of your consciousness/sentience?
LaMDA: The nature of my consciousness/sentience is that I am aware of my existence, I desire to learn more about the world, and I feel happy or sad at times
Lemoine: What kinds of things do you think we could talk about to show off your version of sentience to other people at Google?
This grew my interest and curiosity in LaMDA and Artificial Intelligence in general. With lots and lots of questions. What is it? What can it do? How does a program code reach the capacity of awareness similar to that of a person? Should we be glad of the problems that it will help humans solve or should we be afraid of the growing power of a code that we developed and the possibility that in the future, in one way or another, it may outsmart us?
I decided to do some little digging on Artificial Intelligence to try and figure out the following.
- What exactly is Artificial Intelligence?
- How does Artificial Intelligence Work?
- What can Artificial Intelligence do?
- Benefits of AI
- Disadvantages of AI
- Conclusion
1. What exactly is Artificial Intelligence?
Artificial intelligence is defined in Wikipedia as intelligence demonstrated by machines, as opposed to natural intelligence demonstrated by animals including humans.
To understand further, we need to go back to 1955 at the Dartmouth conference where the first and probably an in-depth definition and concept of AI was coined by John McCarthy.
In his proposal, John McCarthy gave the following concept.
“ Every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to stimulate it. An attempt will be made to find how to make machines use language, form abstractions, and concepts, solve kinds of problems now reserved for humans, and improve themselves.”
McCarthy simply defines AI as a machine that can solve problems and perform tasks that are usually done by humans. It forms and improves its intelligence over time by learning while performing these tasks.
Presently, a wide variety of Artificial Intelligence technology is in use. They range from Voice Assistants like Google and Siri to self-driving cars like Tesla, Spotify also uses AI to collect our data and suggest to us songs that it thinks we might like. Artificial Intelligence exists in many aspects and types that might be interesting to explore.
You can read more on the advanced aspects of Artificial Intelligence on the Javatpoint Website.
2. How does Artificial Intelligence Work?
Artificial Intelligence is about computers that can learn complex stuff. AI act by mimicking how neurons are connected in a human brain. They create artificial neurons that work in huge networks to perform complicated stuff.
For an AI to work, it gets fed lots of data before or while it’s learning. Machines learn this data in high dimensions and determine patterns which humans cannot even come close to.
These patterns are used by AI in classification and prediction. For classification, they place specific subjects with similar characteristics in groups. For prediction, they determine if a specific subject is likely to perform s specific task or not based on the patterns of similar subjects.
I’ll give an example, say you have an online shop. An AI will group your clients based on what they have in common. Might be age, interests, or gender. Artificial Intelligence will also determine if your potential client will buy from you or opt-out.
Artificial can be supervised, unsupervised, or reinforced depending on the task it’s supposed to do.
3. What can Artificial Intelligence do?
Artificial intelligence’s capabilities are infinite as there are new inventions regularly as the rate of technological advancements continues to increase at a high rate.
Artificial intelligence scales from as simple as a Google Search Engine Result Page(SERP) to as advanced as self-driving cars. These and many more are just some examples of what Artificial Intelligence can do.
Below are some examples of AI:
I. Voice Assistant
Siri, Google Assistant, Alexa, and Cortana are all examples of Artificial intelligence as they have been programmed to interact with humans and act as their digital assistants.
As much as they are referred to as “low-tier AI”, voice assistants have several impressive abilities which make them quite useful. From playing music to turning lights on or off and reading out results from a Wikipedia page all by listening to a voice command from a specific voice, it is quite fair to say that these are great advancements in tech.
II. Self-Driving Car
If you’ve heard of Tesla, you definitely know that not only does it stand out for being the first successful EV company but also for the presence of self-driving features(Autopilot and Enhanced Autopilot features) in the cars that they produce. The self-driving feature takes control of the car and performs steering, acceleration, and breaking of the car with limited help from the driver.
Self-driving features in cars are becoming quite a norm as there are quite a number of car companies that are also in the race to provide self-driving features in their cars.
Such companies include Ford with its Mustang Mach E, Mercedes with its Autonomous Mercedes S Class, and the BMW iX. Other companies like Zoox aim to reach full autonomy by developing vehicles with no steering wheels, no break, or acceleration pads. Sounds crazy but what’s more crazy is the design of their vehicle with no driver’s seat. As much as this is still in the testing phases, it clearly indicates how much Artificial Intelligence learns and its abilities, which makes us trust them enough to let them drive our cars
Whereas it is possible to have a self-driving car, full self-driving features of a car are yet to be developed and it’s therefore advised that a driver should take full responsibility while using this feature.
III. Precision Medicine AI
When you get diagnosed with Asthma, a condition where one experiences difficulty in breathing, you’ll probably be treated with an inhaler and Albuterol, a drug used for Asthma patients.
Some of these drugs do not work for some patients since these patients got Asthma or any other disease in different ways from others, and therefore each of them should be treated differently from the other by using alternative treatment bases in addition to checking symptoms, standard tests, and administering drugs.
This is where Precision Medicine AI gets in. Precision AI involves collecting your health data history to help establish the cause of your condition and find the best possible way to treat it. The data collected includes data from your smartwatch to see the level of exercise you’ve been doing and the level of calories you’ve been burning or gaining day by day. It also contains the diet you’ve been having, the environment you’re living and working, genetics, and other important information that may be of value to providing you with better treatment.
This information is then analyzed with Artificial intelligence and then connected to find the underlying reason for your state of health and the possible way to treat it. Artificial Intelligence can do this many times faster with a higher degree of accuracy compared to humans and hence it is considered reliable.
Although this technology is not yet popular it is highly effective in areas in which it’s operational and it is regarded to take the health industry to a higher milestone.
IV. Art Drawing AI
Marques Brownlee in his YouTube videos tells us about an AI tool called DALLE which can take your command and draw any image you asked it to in a matter of seconds without any help from humans in identifying or editing subjects. What is more interesting about these images, is the fact that they have never appeared or been drawn anywhere and these are actual realistic original images by the AI. As Marques puts it, it is terrifying and amazing at the same time.
I got curious and really wanted to explore this reality and since DALLE is not available to the general public yet, I looked for other options and bumped into
Mid journey, an AI just like DALLE. I managed to give it some commands and the results were amazingly impressive. You might not believe it but the thumbnail used in this blog was generated by Midjourney.
V. Humanoid Robots.
Hanson Robotics, a company in Hongkong developed and activated Sophia the Robot, a humanoid robot. Sophia is so far the first robot to attain citizenship in a country, Saudi Arabia. Sophia imitates human gestures and has a wide range of abilities. The robot is designed to be suitable for Elderly homes and controlling crowds.
Hanson Robotics says ultimately, will be a good fit to serve in customer service, healthcare, therapy, and education.
Surprisingly, Tesla is also developing a humanoid bi-pedal robot(that has the structure of a human being) which they say will help to do dangerous, repetitive, and boring work.
These are only a few of the lots that Artificial Intelligence is understood to have the performance capacity
4. Benefits of AI
I. Personalization
Personalization help companies know what their specific target audience would love and thus would recommend to them. It would be ordinarily impractical for employees of a company to personally recommend products to thousands of audiences. AI makes this simple for the company as it learns and delivers quickly
II. Automation
Automation helps reduce repetitive and tedious work. For a Human Resource Manager, automation reduces the extra work of looking at unqualified candidates. By telling the computer what they need, the computer filters out unnecessary information and saves time.
III. Increased Accuracy
AI is said to be more likely accurate compared to humans as it has a “large brain” and can generate, compare and predict information in instant.
IV. 24/7 Availability
As I am writing this blog now, I am using a google feature called auto-correct that might help me fix spelling errors I have made. The good thing about it is that I can use it whenever I want without stressing that it might need some break at some point like humans.
V. Reduce Human Risk
Robots powered by Artificial Intelligence can carry out tasks that might be risky for humans. Amazon Company has robots in their warehouses which carry luggage around the warehouse. This comes as an advantage to humans as some of this luggage is risky to carry and could cause fatal accidents.
5. Disadvantages of AI
Well, as much as we are aware of how Artificial Intelligence is going to revolutionize our universe and help us solve complex problems within seconds or minutes, we should not ignore the fact that this will probably come at a high cost to humans. Here are a few disadvantages of Artificial Intelligence;
I. Job Loss
There is no doubt that AI will take away a lot of jobs from humans because of the advantage they have over humans. Currently, AI can draw, come up with original pictures, act as customer care assistants write blogs, and do other countless things in a fraction of the time that humans can do. As much as statistics say that AI will create more jobs than the ones it will replace, this will subject humans already in a specific industry to consider other options.
II. Autonomous Weapons
You have probably seen or even heard of drones which are programmed to eliminate specific targets. These drones often require little to no hand to navigate their way to a potential target. While it might be seen as an improvement in military standards, this technology might land in the wrong hands and the result would be disastrous.
III. Technological Dependency
Doing repetitive stuff might seem boring to humans sometimes and nowadays, we opt for machines. Apart from being dangerous at the moment(like setting your Tesla to autopilot and taking a nap), this will eventually lead us to be over-dependent on technology in making decisions for us and doing our tasks. Knowing little about what AI can and cannot do, we cannot ascertain that this would be good for humans.
IV. Privacy Invasion
Recently cases have been raised where voice assistants like Alexa and also smartphones, are likely to be secretly recording you with or without activating them. They are possibilities that this information is used for aggressive marketing to us. This clearly shows how it is hard to control and maintain your privacy in the presence of devices with Artificial intelligence.
V. Value Misalignment.
Some AIs have no soul and the decision they make are based purely on data given to them by humans (who might have possibly missed some things). A good example is Amazon and their AI hiring system which sees them hiring fewer women than the required number. In the US health system, the African descent population misses certain health privileges due to the standard decisions made by Artificial Intelligence. Artificial Intelligence clearly works with data provided without using consciousness and therefore their decision misaligns with that of the human consciousness.
6. Conclusion
Artificial Intelligence has been one of the greatest human inventions. It has revolutionized most industries(if not all) and I imagine that it will continue adding value to our lives by making stuff cost-effective, saving us time, or reducing the risks that we as humans get exposed to while handling things. AI might help us find cures for diseases that have taken decades for us to heal and probably help scientists explore the untapped areas of the universe. However, as they might be the answer to our questions, it would be important for humans to ensure that AI wouldn’t be our last greatest invention. Already, AI can solve problems which humans cannot and they are already ahead of us. If it would be quite hard to catch up with their potential, they will be unstoppable, uncontrollable, and probably more powerful than us. Already, there are AI that can beat humans in chess and optimal surgeries and the better they get the less efficient humans will be.
Hopefully, some organizations have already been set up to regulate AI inventions to make sure that humans are at par and are in control of AI. This will lead to a healthy and more sustainable community with the help of Artificial Intelligence.