The AI Era

Preamble:

Year 2011 witnessed the tablet war. Back then, Windows Phone and Blackberry was in action and evolving. The end-user experience was augmented to a new dimension, ease of use and scalability. The concept of tablet computing was gaining traction with a bunch of various devices such as iPad, Galaxy Tab, Blackberry Playbook etc. running on iOS, Android and Blackberry OS respectively.

Years passed and a plethora of other computing devices such as Apple Watch, Android Smartwatch and entertainment based devices and gadgets such as TVs, PlayStations, Xbox came to the fore which raised the bar with modern computing. This was the period when Internet-of-Things started gaining traction. Tech giants locked horns again to build a new ecosystem of connected devices.

The concept of Mobile First approach and Digital platforms started evolving with various App Stores.

The Era of Artificial Intelligence:

So here we are today in 2024 in the era of Artificial Intelligence, and to be specific, Generative Artificial Intelligence. Gen AI is one of the biggest technological breakthrough, revolution and invention which has taken the world by storm ever since the Internet. Be it the creation of an impactful Sales Pitch deck or booking your next travel itinerary or creating a Product Backlog or generating code or converting code from Cobol to Java or composing music or creating multimedia content or creating a recipe... the list goes on... and Gen AI is to the rescue. People from all walks of life can benefit from Gen AI.

The story started in November 2022 when OpenAI launched its flagship product, ChatGPT based on a Large Language Model called "Generative Pre-trained Transformer" i.e. GPT.

Within just two months from launch, ChatGPT amassed 100 million subscribers and set a new record! Interestingly, Gen AI is not a new kid on the block and its history dates back to 1960 as mentioned in the article here when it was introduced in a chatbot. Many decades later in 2014, AI could create multimedia content.

The Gen AI architecture:

A decade back, AI was in action using Siri, Google Assistant and Microsoft Cortana which could perform many tasks for the user and notify them as well for any appointments. This experience was enhanced by Generative AI which has the capability to generate content.

Gen AI is typically based on Foundation Models, Large Language Models and Gen AI models. Foundation Models are the self-supervised and self-learning models which are trained on large amounts of data whereas LLMs can be considered as the subset of these. It is rumoured that GPT-4 Turbo LLM has a whopping 1.76 trillion parameters as mentioned in this interesting IBM Tech YouTube channel video!! In this video, Master Inventor Martin Keen talks about the most important AI trends in 2024.

Now, there is a subtle difference between LLM and GenAI models as most often these terms are used interchangeably. Algolia has a very nice take here. They say that LLM is a class of models while the latter is a type of AI focused on content creation. As per the article published by Elizabeth Bell from Appian here, when she queried to ChatGPT itself about the difference, the bot replied: "Gen AI is the chatterbox at the cocktail party while LLMs are the meticulous librarians! ??".

Tech titans and the AI war:

Various Tech giants such as Amazon, Meta, Google, Microsoft have ventured in the territory of Gen AI and made significant strides in the last 1 year. While OpenAI launched DALLE-3 to generate images, Google launched VideoPoet which generates stunning videos in slow motion. Amazon Bedrock is the Foundation Model used to deploy scalable Gen AI applications. Meta has Llama 2 to show off. Recently Microsoft launched Copilot to generate code, multimedia content and much more. Github Copilot is the most widely used AI Developer tool to perform many tasks such as review Pull Requests, generate unit tests, code comments, documentation etc. Another very impactful and noteworthy AI platform is Anthropic Claude. Claude is a Foundational Model and has various capabilities such as Advanced reasoning, Vision analysis, Code generation and Multilingual Processing.

IBM watsonx has evolved to a large extent in the Gen AI landscape. It is based on IBM Granite, which has a flagship series of Large Language Models based on decoder-only transformer architecture.

Apart from all this, there are some incredibly intriguing developments in Gen AI which is about integrating Gen AI with hardware and chips. In an article published by TechCrunch about Mobile World Congress 2024, Brain.ai demonstrated in the event that AI will be the future of Operating Systems in smartphones. The AI Pin by Humane AI is another such example.

From the many Gen AI use cases, one practical and the most successful success story is about Nvidia AI-designed chips. The chips are incredibly powerful in performance and it propelled the Company to become the third most valuable company in the world. Nvidia is now worth $2 trillion!

Gen AI and Governance:

So, in the backdrop of many such tools and platforms with many features, the concern rises about Ethics, Bias and Hallucination i.e. incorrect responses from the AI models. Gen AI generates content from a large pool of data crawled from the internet. By and large, the content needs to be screened for bias and incorrect responses. Deepfake multimedia content has been sprawling up and so is the plagiarised content. It's paramount that due credit should be given to the rightful owner of the content.

The darker side of Gen AI is about generation of malicious code and cyber attacks. AI in Cyber security could be used to mitigate. Foundation Models and LLMs used in any AI platform play a crucial role here and they should be designed considering all the aforesaid factors. So, it's like AI needed to combat AI.

IBM watsonx.governance is one of the components in its AI platform to handle such issues effectively under the Responsible AI ideology.

Artificial General Intelligence(AGI) and mankind

With all that pep talk about Gen AI, the last but not the least is about AGI. Amazon AWS has elaborated nicely that AGI has the ability to create software like humans. It is perceived that AGI could be a threat to mankind. So, the billion dollar question which looms is whether Gen AI going to replace humans?

Well, in my opinion the answer depends on how we place ourselves in the AI landscape. Gen AI with Prompt Engineering and Tuning is certainly going to augment the jobs and mundane tasks of Business Leaders, Managers, Architects, Developers, Analysts and many such cross functional teams in Enterprises and Organisations. That said, some workgroups which are likely being impacted are Customer care and support and Business Process services.

In a nutshell, the future of mankind with Gen AI platforms is about a transformation in the way an Organisation operates with AI regulations, governance and laws based in that Geography in place. This, in turn, will foster a productive and efficient workforce!