Jeeves or digital Frankenstein? The AI genie came out of the bottle in 2017

Like most things technological, AI is a double-edged sword. Jeeves is what we really want, but we are never too far from Aladdin's Genie. 

God created artificial intelligence (AI) because not all of us can afford to employ a Jeeves — the legendary fictional butler created by P G Wodehouse to help his British master, Bertram Wooster.  

If you have spent 2017 without coming across Alexa, Google Assistant, Siri or Cortana, you may be a technological Rip Van Winkle who needs an algorithmic Jeeves to wake you up from a digital-age slumber. It is time to smell the coffee brewed by these virtual assistants based on increasingly advanced software that mimics human intelligence and is getting to be pervasive. Just as there can be dumb servants and smart butlers, there can be gradations or types of what artificial intelligence or its applications can do for you. You can spend quite some time splitting hairs on concepts close to each other – such as machine learning, deep learning or cognitive computing – but they are all derivatives or subsets of the big thing called AI, which aims to create computer-linked software codes, algorithms, and programs that act much like smart human beings. They watch your behaviour patterns to serve you better, like a friendly spy or ghost. The more you use them, the better they get, like observant butlers.

With social media, internet-linked machines (the Internet of Things) and millions of websites connected to the internet, the big deal is about making sense of the vast amounts of information and then putting it to use. The thing about AI is that it can be fast, like Aladdin's Genie, the mythical precursor to Jeeves from the Arabian Nights tales. The idea is that when something that is already fast is getting to be smarter by the day, we have intelligence on steroids — offering the scope to do plenty of things currently done by human beings at lower costs with higher speed and efficiency. This can get as hip or down-to-earth as you wish.

Cognitive computing is said to be a layer lower than AI but can aggregate data to offer insights and suggestions that cut down months of research to a week or minutes. In nuanced variants, an AI application may tell you which flight to take while a cognitive layer may guide you with rich additional data for you to make decisions yourself.   

Earlier this year, in Mumbai, I met designers Falguni and Shane Peacock who used IBM's supercomputer, Watson, to create Bollywood-inspired fashion design. Watson plumbed decades of data and visuals, approximating to something like having a dozen assistants to throw up creative suggestions. As IBM says: "Watson can think, read and understand natural language such as tweets, text, articles, studies, and reports, and at the same time make contextual sense out of videos and images."

If AI meets fashion at one end, it meets farming at the other. On a pleasant December afternoon, I saw Microsoft experts at the technology giant's Hyderabad campus talk about a collaboration with the International Crop Research Institute for the Semi-Arid Tropics (Icrisat) to help farmers. The two have an AI-based app to help farmers. It uses data about weather conditions, soil quality and other indicators to suggest the best time to sow various crops. The pilot project saw yields increasing from 10 to 30 per cent across crops, says Microsoft, which put its Cortana technologies to help farmers decide on when to sow what. 

AI is getting so better that as a venture capitalist says: "Computer programs will be able to do things that we thought were human-only activities: recognising what's in a picture, telling when someone's going to get mad, summarising documents." Google's annual keynote in 2017 was a lot, if not mostly, about AI getting better.

AI-powered chatbots, which are like a cross between "Frequently Asked Questions" and real human beings, can do elementary customer service stuff. Even Indian companies like HDFC Bank have started employing them to do things like mobile recharges, bill payment and travel booking. A chatbot called Ruuh helps handloom weavers create better patterns and designs. 

Now, look for the threats, mostly to jobs.

Artificial intelligence made way for human acrimony this year when Facebook CEO Mark Zuckerberg and Tesla's Elon Musk fought a billionaire dogfight on Twitter. Musk called AI a threat to human civilisation, while Zuckerberg called his comment pretty irresponsible as he painted a rosy future helped by AI rather than one in which humanity would see an AI-based Frankenstein.

For Indians, there are questions on whether AI and software automation can throw people out of jobs in the $150-billion information technology (IT) services industry that employs four million people. Infosys co-founder N R Narayana Murthy has dismissed the threat scenarios as just hype. However, the man he ousted in a boardroom tussle this year, former Infosys CEO Vishal Sikka, created an AI-based platform, Nia, that is supposed to help IT services run deeper.  

Software service jobs may be safe for now, but it is fair to predict that human beings that act like bots are likely to be less in demand. "Get smarter to outsmart smart bots" may well be a slogan for 2018.

Further, in the land of Aadhaar, we may add that privacy is certainly a concern as virtual assistants and AI-based software platforms stalk us, even if they are doing it for our own good. The cookies that snooped on you in internet browsers may seem like a picnic as Big Brother platforms powered by AI get going.  

Like most things technological, AI is a double-edged sword. Jeeves is what we really want, but we are never too far from Aladdin's Genie, which is not so easy to put back in a bottle.

The author is a senior journalist and editor who has worked for Reuters, Business Standard and Hindustan Times. He is currently an independent media entrepreneur, consultant, and columnist. He is listed among the top 200 Indian influencers on Twitter. He tweets as @madversity