ChatGPT has created barriers to stop malicious content on the platform. However, crooks try to work their way around these to produce phishing emails and malware code. They try to conceal their identity by breaching ChatGPT’s access modes through bots. They use OpenAI’s API to bypass the barriers. They create Telegram bots that use the API. These are promoted in hacking forums. They develop crypto currency payment system. They get people to invest in cryptos that do not exist.

ChatGPT itself can be used as a malware-as-a-service instead of independent hackers. People should be alert and should develop their own AI engines to identify malicious requests, phishing attacks and social engineering. We should increase the cyber education of employees.

Web today consists of 2.35 crore petabytes of data, and the data grows by 70 terabytes per second. Web is thus a vast data store that doubles in size every year. ChatGPT has been trained on internet data up to 2021, and misses data thereafter. A search engine has access to the latest updates on the web.

Computing power required for generative models and search engines differ. Generative models have been trained on massive amount of data and consume considerable computational resources. In its training, there are multiple layers of neural networks, which require huge computing power to process data. Generative model is capable of handling complex natural language queries and generate creative solutions in real time.

Search engines use deterministic algorithms that use pre-existing information to produce results quickly and accurately. The model does not use the same amount of computational resources as a generative model does. Search engines do not generate any new content. They rank the web links on certain factors. In order to up date, these algorithms too require significant computational resources. Besides they handle huge traffic of users. All said and done, being simpler, the algorithms comparatively use less computing power.

Search engines are for accurate results quickly. Generative models are suitable for handling complex open ended queries and provide creative solutions. These two offerings can overlap, but serve different purposes. They can coexist.

ChatGPT uses Natural Language Processing. Its GPT model is called Generative Pre-trained Transformer. The key concept is ‘Pre-trained’. In computer science, in the past two or three years, the algorithms are being pre-trained. In language model, they look at a sequence of words and detect a pattern. They anticipate the set of words that could follow, and make sense. To learn these patterns, they use the data of Wikipedia and collection on the Internet. Thus they have pre-trained algorithms to generate new content based on these patterns. The word Generative in GPT comes from this. ChatGPT has learnt the patterns of human language, common word sequences, and patterns of word usage from data on the Internet.

Microsoft-backed OpenAI’s ChatGPT will soon have its rival from Google called Bard which will be tested among users for feedback. Google also intends to add AI features to Google Search Engine, since Microsoft is too working on the same for Bing. In fact, Microsoft plans to imbue AI into all its products.

It is still not clear how Google will distinguish Bard from ChatGPT. As it is Bard will draw information from Internet. ChatGPT has been fed with information on Internet up to 2021. Google’s LaMDA backs up Bard. LaMDA generates AI powered text. A version of LaMDA requires less computing power. Therefore, it can serve more users. It can be further improved with feedback.

ChatGPT has witnessed explosive growth. It, therefore. could turn away users.

Silicon valley anticipates massive change from generative AI.

China is working on ChatGPT-style project Ernie Bot. It can be imbued in the Chinese Search Engine Baidu.

Many a revolution has been witnessed by the world. The Industrial Revolution percolated to the large parts of the world in 300 years. At the same time, the Internet revolution took hardly 30 years to spread in the world, and the mobile phone revolution took just 15 years and the spread still continues. Airlines acquired first 50 million users in 68 years, telephone in 50 years, TV in 22 years, Twitter in 2 years whereas some Internet applications do so in just some days. ChatGPT has taken the world by storm in a matter of days as so many other spectacular things developed in Bay Area.

Microsoft’s AI chatbot Tay was withdrawn 16 hours after its launch. It was using filthy language and puked venom. Facebook launched Galactica trained on 48 million papers, textbooks, lecture notes, scientific information and encyclopedias in September, 2022. The aim was to better organise scientific knowledge. Days after its launch, Galactica was withdrawn as it spewed suspicious and inaccurate information.

There should be transparency about the data the bot has been trained on, about the algorithms used. There should be accountability. The reference material must be publicly available. Anyone must be able to assess the quality and accuracy.

ChatGPT is a step forward from its sibling model InstructGPT or the GPT-3.5 model. The main distinction is the simplicity of interface. It has been thrown open to public who find it easier and intuitive to use. It has initiated a revolution that Google Search did some 25 years ago.

ChatGPT is a cognitive tool that can dynamically create new content, rather than relying on existing content by searching it. Thus it takes skill to stimulate it conversationally and evoke the most relevant insights out of it.

ChatGPT is limited by its training data and bias in the data. It is necessary to improve the data transparency.

The responses could be verboge, but could have inaccuracies. The predictions relate to future which it tries to align with the text at hand. In addition, it has limited ability to respond to emotions in a conversation.

ChatGPT has the potential to change the business the way it operates and interacts with its customers. ChatGPT is a new entrant, but its previous counterparts which were similar language models like BERT (Bidirectional Encoder Representations from Transformers) and ROBERT (Robustly Optimised BERT) have been in use for quite sometime.

Cognitive bots of the future will disrupt fields like content writing and customer service. They will impact decision making and insight gathering across every domain. They will be powerful digital assistants. We can use time saved by taking their assistance for more productive work.

ChatGPT-4 will soon be introduced in the market using as many as 1 trillion parameters, as against the current 175 billion parameters being used for GPT-3. That will make tool more accurate.

ChatGPT created by OpenAI, a not-for-profit AI research lab is a sensation, since its launch in November 2022. OpenAI comes under OpenAI LP, a for-profit organisation, founded by Musk and Altman in 2015. Microsoft has invested $1 billion in OpenAI and is planning to invest further.

ChatGPT, is a chatbot that answers the queries of the users. The whole question-answer session seems to be a communication with a human. The output of ChatGPT depends on its training — browsing of internet or access to new information. It provides answers based on what it has been trained on.

It is thus not a search engine but AI assistant trained to answer users queries. Yet it has stirred up debates whether it would make Google Search obsolete.

ChatGPT scores over Google in so far as the answers are concerned. ChatGPT provides personalised answers. A search engine provides endless links and the user has to figure out what is relevant for him/her. ChatGPT keeps the whole thing interesting, as it is conversational.

Google declared it as a ‘code red’ threat. Google too is working on similar technology called LaMDA or Language Model for Dialogue Applications system.

Though useful, it is premature to assume that ChatGPT would supersede Google. ChatGPT has been criticised for its incorrect answers. ChatGPT requires constant updating, and it is a challenge. Google Search is based on the advertising revenue model, whereas ChatGPT has still to figure out how it could be monetised. The type of business model adopted will decide how big a threat ChatGPT is to Google.

If Microsoft integrates it with Bing, it will end the debate of ChatGPT disrupting the search engine landscape. In fact, AI will be integrated to search engine models, paving a new way.

ChatGPT is an improved version of ChatGPT-3.5 series models, which were improved versions of GPT-3. GPT stands for Generative Pre-trained Transformer. ChatGPT has been introduced on Nov.30, 2022. The dialogue format makes it possible for ChatGPT to answer follow-up questions, admit its mistakes, challenge incorrect premises and reject inappropriate requests.

ChatGPT has taken internet by storm since its debut. It produces poems, screenplays and essays. In near future, ChatGPT or its improved version will find a lot of practical applications. ChatGPT reached 1 million people as soon as it was launched. By providing better responses to user’s queries, ChatGPT will be a threat to Google Search. Microsoft intends to invest heavily into ChatGPT.

ChatGPT may affect the demand for and employment of knowledge workers.

Is it capable of replacing humans? Its capabilities should not be overestimated. ChatGPT is being developed as a tool which can write like humans. However, it still lacks common sense in the comprehending how the world works both physically and socially.

ChatGPT is here to stay. It will fit into our lives as an assistant or as a tool that augments human capabilities.

ChatGPT was launched last year in November, 2022. It has lit up internet, and gathered its first million users in less than week. It imitates human conversations. It has the potential to supplant professional writers.

The new technology is built on OpenAI’s GPT-3 language model and is powered by AI. The same company has Dall-E image generating model which accepts prompts to synthesize art and other images. Thus AI is getting infused into creative pursuits. OpenAI is already working on GPT-4 model for its natural language processing.

Search engines provide a basic set of links, but this bot ChatGPT is capable of responding to queries in a natural and human like manner, can carry a conversation and can answer follow-up questions. It threatens search engines such as Google.

Microsoft partnered with OpenAI by making an investment $ 1 billion. Microsoft is in further talks with OpenAI to invest as much as $10 billion in it which has been valued at $ 29 billion. Microsoft would be entitled to 75 percent of OpenAI’s profit until it recoups its initial investment. After that threshold, it would retain 49 per cent stake in OpenAI. Other investors who enter with MS would hold 49 per cent between them. The non-profit parent would retain 2 percent. This implies a current valuation of $29 billion. Once the initial $10 billion is recouped in profit sharing, there could be a cap on future profits to be pulled out by the investors. It is highly unusual. Microsoft could add ChatGPT to its Bing search engine. It is being said that MS will add ChatGPT to its MSOffice suite.

Altman, founder of OpenAI, himself wants to make ChatGPT more robust and truthful, and cautions about its premature use. NYC schools have concern for its accuracy, and have banned its students from accessing ChatGPT. ChatGPT lacks the judgement required to distinguish between fact from fake news or fiction.