What is AI? its functioning & misuses.

What is AI?
Before writing about how it is used, it’s advantages and disadvantages i want to explain what actually AI is.
As many of us know AI stands for Artificial Intelligence but it is not like Human Intelligence it is basically is like a program which is capable of using language models, recognizing images, generating stuffs based on previous information and is also capable of improving itself using Machine Learning.
By improving itself if you are thinking of consciousness, you are comparing it with Human Brain and you are wrong, Human brain has many functioning parts and every part has its own role and one of them controls consciousness and other one Intelligence, but AI has only intelligence.
What are Language models?
A question might rise in your brain that if AI has only intelligence how it can chat, recognize, generate and automate stuffs.
First let's talk about how it is capable of chatting but before that we need to understand Language models.
Language Models are basically trained in such a way that it can understand human language and respond according to that, Here Human language doesn’t mean English, Hindi or Urdu here by using term Human Language means how Humans talk or the way of expressing themselves too other.
Again, You might question If machines can understand languages like Hindi, Urdu or English why they need language models. Let’s understand it with an example if someone is saying “It is raining cat & dogs” for humans it means it is raining very heavy but without Language Models computers will not be able to understand it and they will flag it as a grammatical error.
They are used in many places like search engines, Grammar checker and so on. AI use such models to understand our language and use it to respond.
How it recognizes images?
Many images are trained on neural network (A type of machine learning) and AI use them, AI can’t understand what dog is? but what it does it checks for previously trained images and matches them after matching the recognize what is dog.
How AI works?
When you type a prompt AI uses language model to understand your prompt then it searches for pre-programmed actions and according to that they again use language model to respond.
If you are asking Bing copilot “What is CFC?” it will use language model to understand your question and then it will search it on Bing and according to that information and language model it will reply to you with the answer and give you human like chatting Experience.
Same applies for images and for voice it uses services like TTS (Text to speech) and speech to text.
What is the Dark side of AI?
By saying dark side of AI, I mean dark use of AI, I am not going to write ab possible crimes I am going to write about crimes that are already been committed using AI.
Some of them are: -
- Deepfakes.
- Driverless vehicles as weapon.
- Tailored phishing.
Let’s discuss about each of them separately.
Deepfakes
Do you remember image recognition and machine learning mentioned earlier, Deepfake is a technology which uses those technologies combined with AI to create fake images and videos of any person without their consent.
Deepfake is able to swap faces, change expression and synthesize voices.
If You are comparing it with Photoshop edited images it is more dangerous than that here is why: -
- In photoshop a person has to manually edit and apply filters which is a very time-consuming process but Deepfake can do all of these things in a snap.
- Because in photoshop it was manual it was easily detected but it is near to impossible for even a trained eye to detect.
- Normal citizens were aware about that but most of the citizens are not completely literate about AI & Deepfakes and hence it is easy to fool or blackmail them.
For Example,
On November 2021 a 73-year-old man in Kerala lost INR 40,000 to scammer due to deepfake.
The scammer used AI enabled software to call the victim in which scammer’s face and voice matched to the victim’s colleague who was supposedly in urgent need of financial help.
Mostly Deepfake is used for pornography, political and personal purposes.
Driverless vehicles as weapon
As I mentioned above AI can automate devices, someone in Arizona killed a pedestrian using AI droved Uber.
AI can use image recognition and machine learning to target, Identify and kill anyone.
Killer used AI to automate the Uber and fulfilled his intensions.
Tailored phishing
Phishing is a common cyberattack that involves sending fraudulent emails or messages to trick people into revealing their personal or financial information.
But using AI a scammer can use natural language processing to generate human like natural and genuine emails which can be used to trick most of us and if it is tailored for a specific person there is no way to detect it.
For Example,
In 2019, a UK-based energy company was scammed out of $243,000 by an AI voice that mimicked the accent and speech patterns of the company’s CEO.
Opinion of some famous persons
Everyone has different opinions about AI some of them are positive while some of them are negative.
- Elon Musk thinks that AI is a threat to Human Civilization.
- Stephen Hawkins worried that AI may replace humans.
- Sundar Pichai has a positive opinion regarding AI, and he thinks AI is probably the most important thing humans have ever developed and will not destroy us.
My opinion
I agree with Sundar Pichai and i also think AI has the potential of helping humans in every field and will automate simple tasks.
But we should not ignore warnings of other people's we should regulate AI development and usage and we should also not make them able to replicate themselves.
And also, I don’t think AI is going to replace humans but humans using AI can probably replace other humans, but we should use AI wisely.
AI is like a knife, and it can used for both to cut an apple and help you or can be used to murder and second one is dangerous.
There is only one way to stop it and it is to spread awareness about that.
Thank You
Thanks for reading till here it was not all more about AI in other post.