Artificial Intelligence - Corporate Euphoria or Dysphoria?
0
1
0
Everything Everywhere AI at once!
I recently eavesdropped on a conversation between my 4-year-old cat Apollo and the 15-week-old puppy Odin.
Apollo was bragging about how he is using Machine Learning to maximize the crunchies he is getting.
Odin said, “Dude you are ancient! I am using GenAI to parse what Mom is saying and have already tripled the T-R-E-A-T-S!”
Our other 4 year old cat Artemis chimes in ..”Oh please, none of you have a clue what you are talking about! I mean, Apollo, c’mon. You cannot do that without a CNN architecture. And Odin, really? Tripled? Without quantization or hardware accelerators? And neither of you are considering the possibility of hallucination”.
I know. She is a smartass.
All kidding aside, tell me this is not what you are hearing every day, everywhere you go. I mean, when thousands of articles are being written on AI every single day, then the situation is dire, right? According to Statista, as of the beginning of 2023, China produced 76,300 publications in the field of AI between 2016 and 2020, which was the highest amount worldwide. The United States and India followed with 44,400 and 27,000 AI-related publications over that period.
How amazing is the fact that ChatGPT was launched on November 30, 2022. It acquired 1 million users just 5 days after its launch. It has over 180 million users and gets approximately 1.5 billion visits per month.
In less than a year, students are writing pages of apparently decent articles in seconds at such an alarming rate that educators are scrambling to change the curriculum and tests to contain and manage the use of ChatGPT. Lawyers are already getting into big trouble for using ChatGPT for writing briefs with fundamental flaws! I, on the other hand, use ChatGPT very responsibly, only to settle domestic disputes or to impress my motorhead husband with thoughtful comments whether a Porsche Cayenne or a Mercedes E63 AMG wagon is a better dog car!
AI the next big thing?
Is AI a new thing, which showed up as a new virus at the end of 2022? Is AI the next big thing? First, AI is not new and second, it is already bigger than we know!
Artificial Intelligence was founded as an academic discipline in 1956, two years after the death of Alan Turing, the father of AI. Computer Scientists have been continuing research since then. With the progress in cloud computing, AI implementation became a reality and steady progress has been made in the last few years.
The tech giants have been using AI for a long time. How else do you think the insanely personalized ads appear in your search engine or the social media feeds? How do you think your feed itself gets so very tailored to what you think you want to see? Are you aware of the massing computer server farms the tech giants have?
Along came GenAI
Thanks to ChatGPT, AI is magically free! Hold on! So, AI will not be a super special technology only available to government organizations or tech giants? You and I can now use it, and for FREE? Think back to 2003, when we had FREE internet! Did that revolutionize our lives or what? In the same way, ChatGPT has democratized AI in a way that we will not be able to think about our lives without AI going forward.
While ChatGPT 3.5 is free, there is a paid version of ChatGPT Plus for $20 per month. That is similar to what you spend on a dating app!
The democratization of the internet around 2003 has changed our world in ways unimaginable. ChatGPT is being a disruptor in possibly a bigger scale by democratizing AI. Microsoft, Google and Amazon, among the tech giants, have all released drag-and-drop or no-code AI tools. We have no idea how our lives are fundamentally going to change with the democratization of AI.
“Skate where the puck is going to be, not where it has been”, all nice and fine. In the corporate world, people have absolutely no clue where the AI puck is going to be, let alone where the puck is now. Instead, every single middle manager through senior manager is sprinkling AI related terms so liberally that the corporate buzzword bingo games are becoming boring! With little to no understanding of data science, AI or machine learning, they are happily interchanging these words in their effort to win the corporate competition to sound knowledgeable! Even simple decision engines are now relabeled as “AI-Powered”!
Remember those days when folks working on “business intelligence” were gods? Well, now the same people have been repurposed to be “data scientist” and most of them are “AI experts”. Last year, everybody in LinkedIn were sharing how they are getting certified in Cloud computing? Well, this year, everybody is now getting certification in GenAI!
LLM used to be a law degree and NLP stood for Neuro-Linguistic Programming for Psychotherapy, not too long ago. Now, people with very little understanding of anything technology-related, are talking about LLM vs. NLP!
Now that my mother and mother-in-law are also talking about AI, we have officially embarked on the new revolution … everything is #AIPowered.
2024 and our new lists of corporate initiatives
Not to be left behind in the corporate political landscape, most organizations are setting aside 10-ish percentage of their annual CAPEX budget for AI-related initiatives, whatever those may be!
A lot of organizations typically focus on only two things as measures of corporate initiatives: Budget and Timeline. They don’t focus on any other success measures and invariably end up managing their project portfolio penny-wise and pound-foolish. Can you imagine what will happen to them trying to manage AI initiatives in a similar manner?
The first thing to understand is that AI is not a “project”. Even after you complete your on-time on-budget “AI Project”, most probably you will be scratching your head trying to figure out what business value you received and whether you royally messed up spending 10-ish percent corporate budget on a very well executed project with absolutely NO business outcome!
Developing a successful AI Strategy
In my opinion, there are Three Critical Steps for a successful AI strategy:
Understand what AI can do for your business and formulate appropriate use cases.
Tech giants have established AI maturity models and they are way, way, way ahead of the curve. You just need to choose an appropriate starting point which is tailored to your own businesses. That’s it. What is AI and how will it help your business?
As I said, most people don’t even understand what the “AI puck” looks like let alone try to skate to it! So, it is important to take baby steps. Start simply. Think about your immediate opportunities for task automation. Focus on a few of those with crystal clarity of desired outcome. Create a clear vision of “To be” State and also figure out how you will measure your outcome. Do not try to solve complex partial derivatives equations, just stick to simple algebra as a starting point.
Data, Data, Data
A long time ago, in a galaxy far far away, the concept of a data warehouse started. Well, it may not have been that dramatic! However, by 2010, most large and mid-size organizations had established a “data strategy” and started executing. Along with the evolution of data techniques, the strategy continued to mature. Data warehouse, marts, big data … we were evolving at a steady state. And then came the massive acceleration with cloud computing.
The data-mature organizations started initiatives to move data into the cloud and unleash amazing insights. Big data was indeed a big thing!
However, not all organizations are data-mature, especially the ones which depend heavily on external sources and systems for data. This is true for organizations which outsource key aspects of business to other companies and become consumer of data. It is not easy for these organizations to clearly articulate “what is my core data”. More often than not, they lack a data strategy.
It is critical to be brutally honest about the state of your data affair. If you don’t have a robust data platform, then that is the first thing you need to focus on. Don’t hide behind the buzzwords of “data lakes” and stuff like that. It will only lead to failure. If you are behind the eight ball, this is your opportunity to rectify. Instead of the 10-ish percent corporate budget for AI initiatives, spend that on a data platform. With the latest technology and techniques, it is no longer “boiling the ocean”, as many organizations are fearful, and stay paralyzed to undertake initiatives to build a robust data platform.
Now, instead of AI, why the heck am I talking about data?
For AI models to work, you need a lot of data, and not just a lot of bad data. Good and clean data. The biggest asset tech giants have is data. Here are a few examples:
● Google handles over 3.5 billion searches daily
● WhatsApp users exchange 65 billion messages daily
● Internet users generate 2.5 quintillion bytes of data daily
● Facebook generates over 4 petabytes of data daily
● Amazon holds over 160 exabytes of customer data
Tech giants have been perfecting AI models with this insanely huge amount of data, for years!
Without having a robust data platform to provide solid data for a model to learn, what will end up happening is “garbage in garbage out”! A model maybe really awesome, however, the insights may be really flawed if the underlying data is not solid.
So, first things first. Don’t hide behind any excuse. Focus on building a robust data platform. If you have a solid data platform, then focus on data quality. AI will not work without a lot of good and clean data.
don’t worry about boiling the ocean. You don’t need a pristine store of structured data. LLM’s work well with unstructured data. Focus on data integrity and data quality, and data completeness
Remember to Iterate and “train”
Going back to the prior discussion on “AI Projects”, it is critical to understand that AI is non-linear. It is not a one and done project. It requires iterative maturity. This ensures high quality output tailored specifically to your use cases.
ChatGPT took 4 years of human feedback to build.
Unless we understand that the models need “training” and tuning and adjustments, before you know, AI initiatives will not deliver desired results and you will be sorely disappointed by the outcome.
**RAG – how to optimize training.
Happy AI’ing!
** if done wring it will be dysphoria , bit AI is not optional.
In summary, AI is not a new thing, but democratization is the revolution which will impact us in ways we can’t even fathom. Our lives will fundamentally change in the next 5 years and a whole lot of task-based and rule-based activities and jobs will be eliminated.
This is not a bad thing! This is actually a great thing! While we all get blue in the face debating the ethics of AI, we will finally start appreciating the VALUE of human work, as opposed to the mundane focus on, yes, you got it, Budget and Timeline!
Time has come for us to stop talking and start doing! Let us get focused on the most important thing which AI cannot do … focus on business strategy! And then let’s leverage AI as much as possible to remove friction points to automate task-based activities and execute, execute, execute!
Related Posts
