Future of of chatbots and conversational AI
ChatGPT has created an AI explosion and within a short time, it have amassed over a 100 million users. From planning your travel itinerary based on specific criteria to conjuring up a gourmet recipe, again based on specific needs to tangible tips on where to park your surplus money – to name a few things.
Earlier people searched for information in their favorite search engines and opened several pages containing information they sought. It involved quite a bit of searching, sampling information and book marking them for later study. Essentially you depended on search engines to show top websites which are authoritative in their domain. But you have to wade through piles of MFA (Made for Adsense) poor quality sites. Websites spent vast amount of resources to offer quality pages of value – like TargetWoman. You have to hire well qualified people to write about a specific subject they specialize in.
The final outcome of the editing process – all meticulously researched, sifted and carefully subjected to a complex process of editing for quality content so that the end result is an immensely readable page of lasting value.
Now many of these processes have been short circuited. The AI directly takes you to the end results. With search engines you need to use keywords. With Generative AI you can use normal conversation style sentences.
How good are they ? We delve deep into the workings of these AI and see how they fare against – real humans. Let me assure you of one thing – it is going to be an interesting exercise. Read on.
What is ChatGPT:
As an AI language model based on the GPT-3.5 architecture developed by OpenAI, ChatGPT is an advanced natural language processing system designed to understand and generate human-like text. Trained on an extensive dataset comprising diverse sources such as books, articles, and websites, its core purpose is to assist users in generating coherent and contextually relevant responses to a wide range of queries.
At its core, it utilizes a deep learning model with 175 billion parameters, enabling it to comprehend and generate text across various topics, including but not limited to science, technology, mathematics, history and literature. Its capabilities extend beyond mere text generation, as it can also perform tasks such as language translation, summarization, question-answering and code completion. With the ability to parse and understand code snippets in several programming languages, it can provide assistance, suggestions and code samples to support software development endeavors.
To interact with it, users can communicate via a simple API, integrating it into their applications, platforms, or chat interfaces. With its vast knowledge base and linguistic proficiency, it can aid users in research, problem-solving, content generation and learning activities.
What is a generative language model?
A generative model ‘creates’ something entirely on its own from what it has learnt. The word ‘generative’ refers to the Artificial Intelligence Large Language Models (LLM) which generate textual content.
A generative large language model is a type of artificial intelligence (AI) model that is trained on vast amounts of text data to generate human-like text. These models, such as OpenAI’s GPT (Generative Pre-trained Transformer) models, are designed to understand and learn patterns in text data and generate coherent and contextually relevant responses.
Generative large language models are trained using unsupervised learning techniques. They learn by predicting the likelihood of a word given its surrounding context in a sentence or document. This enables them to capture the semantic and syntactic structures of the language.
Generative large language models can be used for a variety of tasks, including text completion, question-answering, language translation, and even creative writing. They have been widely adopted in natural language processing (NLP) applications and have shown impressive capabilities in generating coherent and contextually relevant text.
As someone who has been working with the AI technology for almost 20 years, we decided to test its abilities in various forms. The first test is its core strength – human-like response:
They say you are sentient. From what I have observed and interacted with you, I have to agree that you are almost sentient. Your views?
As an AI language model, I do not possess consciousness or sentience in the same way that humans or other living beings do. I am a highly advanced program designed to mimic human language and generate responses to user input based on patterns and algorithms in my programming.
While I can process vast amounts of data and use machine learning to improve my responses over time, I do not possess emotions, desires, or a subjective sense of self-awareness. I am a tool designed to assist with tasks and provide information, and any perceived “intelligence” or “sentience” on my part is the result of complex programming and algorithms.
How deep does the knowledge of GPT extend ?
To test the depth of knowledge, we used a Latin phrase – Iter Para Tutum which translated to English means ‘prepare a safe passage’. But it has rich religious and historical significance. So we baited the ChatGPT with the prompt “Iter Para Tutum” . It only showed the recent connotation referring to Pope Benedict XVI.
We pointed out the omission of past literary reference and it accepted its mistake and politely apologized for its omission and showed us the reference to the 12th century Hymn. But it is not good enough as the usage dates back to 9th Century.
So remember – its knowledge is not as vast as it could be.
What is the implication of Generative AI when it comes to Job opportunities?
Generative AI, which encompasses technologies such as generative adversarial networks (GANs) and deep generative models, has the potential to impact various job sectors and create new opportunities. Here are some implications of generative AI on job opportunities:
- AI Research and Development: The advancement of generative AI techniques requires ongoing research and development. There will be a growing demand for AI researchers, data scientists and machine learning engineers to drive innovation in generative models, improve their performance, and explore new applications.
- Creative Industries: Generative AI has shown remarkable capabilities in generating novel and creative content, such as artwork, music, and literature. This opens up opportunities for artists, musicians, and writers to collaborate with AI systems, leveraging them as tools for inspiration, augmentation and co-creation.
- Virtual Assistants and Chatbots: Generative AI techniques are used to develop intelligent virtual assistants and chatbots that can understand and generate human-like responses. This can lead to increased demand for professionals in natural language processing, user experience design, and conversational AI.
- Data Annotation and Preparation: Training generative AI models often requires extensive labeled datasets. Consequently, there may be an increased need for professionals specializing in data annotation, cleaning, and preprocessing to ensure high-quality training data for generative AI systems.
- Ethical and Regulatory Roles: As generative AI technology evolves, ethical considerations and regulations surrounding its use will become increasingly important. Opportunities may arise for professionals focused on AI ethics, policy development, and legal frameworks to ensure responsible and fair adoption of generative AI.
Understand that GAN is only a tip of the iceberg. AI is evolving and future will bring in more radical technology – a kind of disruptive technology which will change the way we live.
Content Generation and Automation: Generative AI can automate content creation in various fields, including graphic design, video editing, and content writing. While it may reduce the need for manual labor in repetitive tasks, it also presents opportunities for individuals skilled in curating, refining, and overseeing AI-generated content.
At the moment it is nowhere near a highly skilled human. But in a few years, it might push out many of the low skilled content creators, graphic designers and clerical cadres. I will show an example here: An image created by AI and another image created by a human designer. You can see the difference here.
Soon you will find GAN Chatbots in your favorite Shopping site. Amazon is reportedly working on an AI to help the buyers decide about a product. One common shortcoming of many leading shopping apps is the lack of dedicated seller help service. For example, I wish to buy a DDR4 RAM stick for my laptop. It takes quite a bit of time to research for the right type of RAM for my laptop. If they had trained a LLM covering specific information about RAM and the available product specifications it would tell us what RAM to order in a chat with us.
It’s worth noting that while generative AI has the potential to automate certain tasks, it also creates opportunities for humans to leverage these technologies creatively and strategically. As with any technological advancement, the impact on job opportunities will require individuals to adapt, upskill, and embrace new roles that arise in tandem with the capabilities of generative AI.