Artificial Intelligence – coming of age?

The?Babbage Difference Engine, a mechanical computing device was designed by Charles Babbage in 1822 But it was a 120 years later, in 1942 that the first modern electronic digital computer, the?Atanasoff-Berry Computer?(ABC) was designed by Prof. John Vincent Atanasoff and his student Clifford Berry at the Iowa State University. From there the evolution of modern computers kept accelerating. The ABC was never patented hence the creators of the?ENIAC?(Electronic Numerator Integrator and Computer) in 1947 took out a patent and claimed to be the inventors of the digital computer (the ENIAC patent was thrown out by a judge in 1973).

While these developments were focused on computing or processing capability, there was also a parallel effort in exploring the ability of machines to learn and develop intelligence similar to that of humans. In 1950, Alan Turing (yes, the same mathematician who cracked the ENIGMA code during the second world war) pioneered machine learning. In his publication “Computing Machinery and Intelligence”, he proposed a test to check a machine’s ability to exhibit intelligent behavior equivalent to human intelligence (called the?Turing Test). In the years that followed, there has been sufficient research to demonstrate that the imitation game in the Turing Test is not infallible i.e., the ability of the test to determine a machine response can be compromised. Of course, this is also linked to the development of vastly superior computing resources and AI and ML capabilities in the modern world.

The term?Artificial Intelligence?was coined in 1956, by John McCarthy at the Dartmouth Conference. This was also the period when computer languages like FORTRAN, COBOL and LISP were invented.

ELIZA, the first chatbot (although this term was created much later) was created in 1966 by Prof. Joseph Weizenbaum of MIT. ELIZA was designed to imitate a therapist, asking open ended questions as well as follow ups based on the user response!

The emergence of?Expert Systems?commenced in 1965 with the design of?Dendral, by Edward Feigenbaum and Joshua Lederberg of Stanford University, to analyze chemical compounds. The expert system relies on a knowledge base and an inference engine to perform feats of intelligence. The knowledge base is an organized collection of facts acquired from human experts and represented in an “if-then” format of rules. The inference engine interprets and evaluates the facts in the knowledge base to come up with answers. Expert systems incorporate heuristic rules (or “rules of thumb”) so increase learning from experience. These systems are used today as aids to human experts (rather than as replacements) in areas like credit evaluation, fraud detection, medical diagnosis etc.

From the late 1990’s there were more examples of intelligent agents coming to the mainstream. The celebrated event was?IBM Deep Blue?defeating Chess World Champion Gary Kasparov.

Go: An ancient Chinese board game played on a 19x19 grid and has legal board positions computed to 2.1 x 10^170?– a number vastly greater than the number of atoms in the observable universe! Google owned company DeepMind created the AlphaGo AI system that defeated Lee Se-Dol, 4-1. Lee was a former world champion of Go. In 2017, DeepMind created AlphaGo Zero which learnt by playing itself (AlphaGo in comparison learnt by studying a dataset of 100,000 human games). AlphaGo Zero then defeated its predecessor 100-0!!

We then started seeing the emergence of AI and ML in use cases ranging from vacuum cleaners (Roomba) to entertainment (Netflix?recommendations) and social media (with use of AI by the likes of?Facebook?and?Twitter). The use of navigation tools like Google Maps is now ubiquitous (when was the last time you stopped to ask a stranger for directions?). Commercial use was leveraged by ride-hailing services like?Uber,?Grab,?Ola?and?Lyft.

Note: The famous London Black Taxi drivers have to pass a test called The Knowledge, to show that they know the name and location of every street and the shortest or quickest route to it, within a 6-mile radius of Charing Cross!

In the popular realm of Hollywood films, there has historically been a characterization of AI based personas including

·??????HAL 9000, the fictional AI character and main antagonist of Arthur C Clarke’s Space Odyssey series. HAL (Heuristically programmed Algorithmic computer) appeared in the 1968 film 2001: A Space Odyssey.?

·??????Data, a synthetic life form with artificial intelligence, who is self-aware, sapient and sentient is an anatomically fully functional male android serving as the chief operations officer aboard the Starship USS Enterprise-D. He appears in the TV series Star Trek: The Next Generation in 1987.?

·??????J.A.R.V.I.S.?(Just a rather very intelligent system) is an AI platform serving Tony Stark (aka Ironman) who made his debut in the 2008 film Ironman. Jarvis’ main role as Stark’s assistant is to manage all the internal systems of Stark’s buildings and the Ironman suits.

In the real world, we now have sophisticated digital assistants including:

·??????Alexa?– Amazon’s cloud-based voice service available on more than 100 million devices from Amazon and third-party manufacturers. Alexa provides an intuitive way to interact with every-day use technology. Amazon also offers APIs and development toolkits for building solutions with Alexa.

·??????Siri?– Apple’s voice-controlled personal assistant who first appeared on iPhone 4S and is now available on iPad, Apple Watch, AirPods, HomePod and the MacOS. Siri provides a seamless way to interact with the device.

·??????Google Assistant?– Google’s version of the digital assistant offers voice control over your device, search and task assistance. It can be a sophisticated digital concierge to fulfil your needs from booking a table at a restaurant to providing real-time translations.

·??????Cortana?– Microsoft’s digital assistant who had features that were provided much later by Alexa, Siri and Google Assistant slowly became sidelined once the Windows Phone went extinct. Cortana was finally killed by Microsoft in early 2021.

While there is a justified fear that improved AI based solutions will eliminate jobs in the future, it is also equally true that a completely new set of opportunities will present itself to the workforce of the future. Consider what AI based navigation has done to the transport industry – how many more people across the world now drive taxis (as part of ride hailing platforms) to gain employment (or to supplement income).?

Sundar Pichai?who was recently in India said in an interview that he “sees AI as a tool that can help humans do many of their tasks better”. AI programs can help chess players improve their game. AI assistants can help with writing, creating art. He concluded “I think humans can do it better with AI”.

The latest burst of interest in AI has been triggered by two interesting projects. In July 2022, deep learning model?Dall-E 2?was launched in Beta phase. Developed by?Open AI, this tool develops digital images from natural language descriptions (or prompts).?

Open AI, an artificial research laboratory was set up in 2015 in San Francisco by Sam Altman, Elon Musk, Reid Hoffman, Peter Thiel, AWS, Infosys and others who pledged over USD 1billion to the venture.?

In November 2022, Open AI launched?ChatGPT?(Generative Pre-trained Transformer), a chatbot, as a prototype. This tool uses supervised learning as well as reinforcement learning with human trainers playing a role.

The excitement generated by the capabilities of ChatGPT is immense. The tool can compose poetry, write essays, answer queries that can be posed by students or customers of any product or service. It can potentially create legal documents as well as literature.

However, there is a health warning that OpenAI has given about ChatGPT stating “it may occasionally generate incorrect information” and “produce harmful instructions or biased content”. In fact, Meta recently removed the preview of its chatbot Galactica because of the incorrect/ biased information it was propagating. Microsoft released Tay an AI chatbot on 23 March, 206. However, when it began to post inflammatory and offensive tweets through its twitter account, Microsoft shut down the service just 16 hours after launch! Tay learnt from its interactions with trolls who fed it biased and offensive information.

Clearly businesses have to be tuned in to the opportunities that various AI platforms can create as well as the threat posed in terms of changing the demand dynamics (apart from the dangers of developing biases with uncontrolled interaction). A recent example relates to the emergence of ChatGPT. This is an enormous technological breakthrough. According to media reports, Google’s management has declared a “Code Red”, fearing that this is an event that can upend their business model. Google has its own version of the intelligent chatbot called LaMDA (Language Model for Dialogue Applications). But the big challenge is that these chatbots are not (yet) suited for delivering digital ads! It was these ads that accounted for almost 80% of the company’s revenue last year.

But it’s also worth exploring missteps by large corporations when it comes to translating AI use cases into realizable revenue opportunities – read this excellent?article ?in the New York Times from July 2021 on what happened with IBM’s Watson.

Continued research in the field of AI shows that we have not yet reached the level of sophistication of the human mind. The current AI models which seem to perform at a human level, do so in a very non-human way and can fail when faced with a completely unanticipated situation. Clearly the level of testing needs to be rigorous to cover ALL possible scenarios and not just the most common ones. When dealing with critical safety systems, a single error can be costly in terms of the dangers it can pose (e.g., In self-driving cars).

It is difficult to predict the exact future of AI, as it depends on various factors such as advances in technology, changes in societal needs and values, and the actions of individual researchers and organizations. However, it is likely that AI will continue to play an increasingly important role in many areas of our lives, including healthcare, transportation, education, and entertainment. It is also possible that AI will be used to address a wide range of global challenges, such as climate change and poverty. It is important for society to carefully consider the potential consequences of AI and ensure that its development is guided by ethical principles.

PS: this last paragraph above was a response by ChatGPT to the question “tell me about the future of AI in a paragraph under 100 words”

Graphics:

1.?????DALL·E 2022-12-28 13.29.38 - a masters vision of artificial intelligence

2.?????DALL·E 2022-12-28 13.29.01 - painting of a human talking to a computer by Dali

3.?????DALL·E 2022-12-28 13.27.16 - painting of a robot??playing chess by van gogh

________________________________________________________________________?

I hope that you enjoyed The Lateral View!?Look forward to connecting with you!?Every fortnight, I'll share my perspective on topics relating to technology, banking, insurance, capital markets, financial services, leadership etc. To make sure you don't miss an issue, if you haven't subscribed yet, just click the "Subscribe" button in the upper right corner above.?

-Shrinath?

Shrinath Bolloju ?is an independent management advisor. He has spent over 30 years in the banking and securities services industry and has worked in technology, operations and run a business, in various geographies across Asia, Europe and the Americas.

DALL·E 2022-12-28 13.29.38 - a masters vision of artificial intelligenceDALL·E 2022-12-28 13.29.38 - a masters vision of artificial intelligence