ChatGPT: The chatbot revolution is here

Paving the way for advancements in Artificial Intelligence

Data Science, Machine Learning, Neural Networks, etc. are no longer new or novel technologies. In recent years, these have rapidly evolved and have been integrated into many different industries and applications- be it the use of ML algorithms in healthcare to analyze medical data or the use of data science techniques in finance to predict market trends and identify investment options, AI is definitely the 'new normal'.

Its integration into our daily lives has enhanced our ability to process and examine large amounts of data. These technologies have become essential tools for solving complex problems and making informed decisions.

Transforming the whole arena of Generative AI

As we progress, advancements in this field are moving towards Generative AI- Deep Learning is being put to use to generate images, text, and even music! The best part about this is that the generated pieces of data are amazing. They not only give detailed responses based on the prompts received but sometimes surprise the user with outputs that were not even expected. Professing this ideology, Open AI, an organization working towards developing tools for Generative AI has really blown the whole industry off- earlier it was DALL·E 2, an image generator and now it has come up with yet another groundbreaking technology, ChatGPT.?

ChatGPT has pretty much cracked the language structures and caters to use cases like answering questions, writing dialogues and essays, summarising articles, coming up with catchy slogans, generating computer code, and whatnot. Let's delve into it.

ChatGPT: Changing the game of Natural Language Processing

It has observed an accelerating number of users using this tool for purposes ranging from generating code to having knotty responses to the prompts. While its ability to come up with dynamic replies raises questions of unemployment, one thing we all can agree upon is that the chatbot is powerful and useful indeed!

LET'S EXPLORE THIS!

I used it for generating a short story, asked it to behave like an Elephant, and gathered its opinions on sensitive topics and it just pulled this off better than any other platform I have used to date. Refer to the slideshow to see its cool responses.

ChatGPT vs Others: Drawing a comparison

The picture depicts how vast the database of ChatGPT is- other chatbots came up with abrupt responses but ChatGPT despite being in its 'research review phase' gives more detailed and 'ethical' responses. Moreover, if it is not sure about a certain topic, it does apologize for not having a concrete opinion about it and comes up with a diplomatic answer for the same.

Analyzing the bias and ethical factor

These snippets are full of positives-

• My experience with using other AI chatbots is one that reflects gender bias specifically. I recall when I was using a chatbot and asked it to develop a narrative as a technician- even after running the same question repeatedly, it was limiting its image of a technician to the 'males-squad'. It kept on using Mr. Technician or typical names of boys. But the way ChatGPT develops its responses, clearly shows how well the bias and ethical factor has been dealt with. Yes, we can conclude that it is more politically correct than any of our politicians.
• A factor to worry about in some of the cases is the chatbot's tendency to repeat a set of words more often than required. Also, when I asked it to give me some information about Virat Kohli, its response was- "Virat Kohli is an Indian cricketer and the current captain of the Indian national team. He is considered one of the best batsmen in the world and is known for his aggressive and attacking style of play. Kohli made his international debut in 2008 and has since become one of the most successful Indian cricketers of all time, setting numerous records and winning numerous awards. He is married to Bollywood actress Anushka Sharma." This paragraph holds a wrong piece of information in its initial lines that have been highlighted.
• This is not the first attempt to build something transformative in the world of chatbots, several other tech giants like Meta and Microsoft have tried their luck but failed after a short period of time. In 2016, when Microsoft launched Tay, within 24 hours the whole functioning went wrong because Twitter users embedded misogynistic notions into it. Likewise, Meta couldn't survive because BlenderBot 3 was spreading racist remarks and false information, such as Donald Trump winning the 2020 elections, according to Mashable.
• To avoid these disastrous consequences, OpenAI set up Moderation API to make sure that language doesn't go against OpenAI's content policy that thwarts unsafe, inappropriate, and illegal information to get the green signal. However, the OpenAI team admits that flaws might still reside here and there that will be rectified over a course of time.

Using it as a Search Engine

Many people are raising the question of using this tool as a substitute for search engines. While I agree with the fact that it puts everything related to your question in one place unlike the search engine wherein you have to go through every single link provided which consumes more time but I believe that a search engine provides you with a plethora of information and that too with minute details in the form of various media (texts, youtube videos, images, research papers) that a chatbot cannot do.

Adding to this, an integration of this in search engines would be cool like in the case of Designer by Microsoft- Microsoft is launching Designer, a tool that creates designs for graphics, flyers, posters, etc. In October, Microsoft and OpenAI announced that DALL-E 2 will be implemented into the program. Microsoft is also integrating DALL-E 2 into Bing and Microsoft Edge which will give users the opportunity to create their own images if they are unable to find something related to their results.

The Progression

Sam Altman, CEO of OpenAI commented, "The chatbot technology is transformative and has great potential. We shall harness it to the fullest and this is just the beginning."

• While this tool is just in its initial phase, this is a good sign. As of now, the users can sign up and use it for free but there are chances of monetization (cost per chat) taking place. Considering its heavy computational power needs, if monetized, it will be out of the range for a lot of sections of society.
• OpenAI has not put any spotlight on how the algorithm actually works but one thing that we know is that the outputs will be influenced by the training texts it receives in its review phase. Hence, it is our responsibility to harness the power of AI tools carefully and not propagate sexism, racism, and hatred in any of our prompts.