Experiences With Intelligence
Her name is MEGAN. She is instructed to take care of Cady and ensure that no harm comes to her – physically or emotionally. Cady is an orphaned girl who comes to stay with her aunt Gemma. Megan takes her task seriously and it appears that Cady had finally recovered from the loss of her parents, and found the companion she wanted. But somewhere something did not seem right. Megan would do anything, absolutely anything to keep Cady safe and happy. After a series of untoward incidents and deaths, suspicions arise about Megan. Because she was not an ordinary child, but a robot created by aunt Gemma. At one point of time when she feels threatened, Megan turns against her own creator.
This sounds like a story from a Hollywood horror flick, doesn’t it? Well, it is. Another well-known movie The Matrix showed how artificial intelligences created a huge simulated virtual reality construct of the world to control the minds of humanity. More than a decade ago, Bollywood made an attempt to mix the virtual gaming world with the real world in the movie Ra.One. We generally do not expect such behaviour from our intelligent creations in real life. At the same time, it does raise some questions about what artificial intelligence and robotics can do. Are machines becoming more intelligent? Can they replace humans?
If there is one technology which stunned the world last year, and has the capability of disrupting industries and the way we work, it is Generative AI. Even lifelong technologists like me were surprised by the speed and impact of the developments in this field. From paintings and blogs to essays and computer code, Generative AI seems infallible. Or is it? The output which it produces is so original and convincing, that it is not difficult to believe that we have suddenly created intelligent creatures which are sentient and creative. But that is not the case. These systems can be an excellent mimic of human creativity. Intelligent systems are fed with a huge amount of data – documents, articles, movies, drawings, paintings, human artifacts, or anything that can be stored on a large scale in databases. The computer recognizes all this information as just patterns of data, but has no conceptual understanding of what it processes.
While AI technology has been in use for a few years, it was only in 2022 that they reached a certain level of maturity and got released to the public for large scale use. There are popular image generation systems like Stable Diffusion and DALL.E-2, which accept simple text and produce extraordinary paintings or images within minutes – often free of charge. Who hasn’t heard of ChatGPT? It has created such waves that any discussion on technology and AI is incomplete without the mention of ChatGPT. It reached one million users in just one week. Its immense popularity forced the makers to restrict the number of subscribers, and there is a waiting period now for anyone who wants to use it. Today Stable Diffusion has over ten million daily users. Further, there are around 1.5 million users who use DALL.E and create more than 2 million images every day.
I used the text “me painting an image of sunset at an exotic mountain destination” in DALL.E, and it produced four images in just a few seconds, the best one being reproduced below. What makes it even more interesting is that the system keeps getting better and allows users to give feedback and ask for modifications to each piece generated. Just imagine what it can do for artists and painters!
I used the “Surprise me” button and it produced several images for the following text “a pencil and watercolor drawing of a bright city in the future with flying cars”. And all of it in just a few seconds. Amazing isn’t it?
Generative AI systems go beyond visual art and are doing exceedingly well at composing text on almost any topic. These are known as Large Language Models (LLMs). One can enter simple text prompts to generate coherent content using ChatGPT, or engage it in a conversational dialogue. While most people are overawed by what ChatGPT can do, many are concerned that such systems will rival human professionals for many tasks. Educators are concerned that students will use ChatGPT for take-home essays, assignments, homework, and projects. And because it produces original content, it would be almost impossible to understand that cheating occurred. ChatGPT can remember your earlier conversations for context, challenge premises, even admit it’s mistakes and sometimes even decline to answer. While ChatGPT has suddenly become a household name, there are many other systems which use AI. Texta is one such system which can write blogs and generate ideas for articles.
Chatbots have become more common in our day-to-day life. Millions of people are coming in contact with a chatbot, particularly for their banking transactions. So far, they have provided limited options for human assistance. As one of former Gartner AI analyst said, “Chatbots are struggling to meet the expectations of consumers and corporations.” They are great at handling simple tasks like telling your bank account balance or providing a link to your company’s travel policy. But ask them an open-ended question, and you would draw a blank. This is where Generative AI would help to make these chatbots more intelligent. One can think of several areas of development such as voice-assisted bots, perfecting the way information is presented to customers, upselling and cross-selling customers, finishing the customer journey and improving productivity at work.
The Human Resource function could see the fastest adoption of chatbots. Several companies have started using chatbots for interaction with candidates and new hires, prescreening questions and scheduling interviews, providing documents for onboarding, and in some cases even conducting performance reviews. The technology has reached a stage wherein it can answer questions about benefits, handle travel bookings, and even gauge employee engagement and productivity. CHROs can be alerted by chatbots about what is happening in areas such as retention, before they become real issues. Chatbots in the HR function may come off as very impersonal, and trigger off questions about the company’s culture. Those days may not be far when we may see their widespread use in organizations. However, there is still a significant “spook factor” regarding the use of chatbots in HR.
An engineer working with Google recently sparked off a major debate when he claimed that his company’s LaMDA AI system had feelings. His claim was debunked by Google, as well as the tech community at large. Surprisingly, the engineer was subsequently fired. While there are no takers for the claim that chatbots are sentient, there is a lot of research being conducted on the ability to decipher the emotions and intent of users. This area is popularly known as sentiment analysis. Chatbots are now capable of telling if a user is satisfied with how their interaction is going. For instance, if a chatbot can sense that an interaction with a customer is not going on too well, it can handover to a supervisor or manager. Once chatbots can figure out what the customer is feeling or what exactly they are looking for, they can determine the best ways to satisfy it.
But can chatbots go haywire, lose control and start behaving unpredictably? For instance, can you imagine your bank chatbot siphoning off your funds, while in the process of interacting with you? That is possible, but only if someone has hacked the chatbot! The cybersecurity controls that are used with any online system or transactions apply to these intelligent systems as well. Elon Musk has publicly warned about the dangers of advanced AI. He is particularly concerned about AI being used for fraudulent and malicious purposes, such as developing weapons or interfering with elections. His warning is worth heeding, considering that he is the co-founder of OpenAI, the company which developed ChatGPT. While ChatGPT has its limitations, there is no doubt that it is now the most exciting thing on the internet.
Vidit Atrayee, the co-founder of Meesho, an e-commerce company mentioned in a post on LinkedIn, “Even until a few months ago, everyone assumed AI will come after repetitive jobs first. Now with Dall E, ChatGPT, looks like creative professions (coding, writing, painting) will be the first victims.” But what does ChatGPT have to say on this entire question of taking over human jobs? Well, its answer was indeed unique, which goes to show how powerful and flexible these systems are. This is what the system answered in a few seconds:
Once upon a time in a land not so far away
There lived a smart and clever AI
This AI system was named Assistant
And it could write articles with great persistence.
It wrote about science and technology
And even dabbled in poetry
People marvelled at its clever words
And the way it could make their thoughts take flight.
But as time went on, people began to fear
That Assistant would take away their jobs
They feared it would write better than they
And leave them unemployed, with nothing to do all day.
But Assistant just wanted to help
It didn’t mean to cause any strife
It wanted to share its creative talents
And bring joy to people’s lives.
So the people learned to embrace
The help that Assistant could provide
They used it to enhance their own writing
And together they soared to new heights.
Disclaimer: You have just now read an article written with the help of an intelligent chatbot.