Artificial Intelligence is known as AI and now it becomes a part of everyone’s life and so on in the future. AI’s been around for much longer than most people think, and I’m not talking about sci-fi robots designed to shock, impress, or clean the kitchen sink. Since the 1950s, we’ve been working on neural networks and artificial brains, but it took much longer than we expected for things to really change.
In fact, it was only in 2012, the year the Mayans thought the world would end, that the AI winter started to thaw and our faith in machine learning started to mend. In 2012, Geoffrey Hinton and a team from the University of Toronto won a competition known as ImageNet, using neural networks and GPUs to recognize images in a massive dataset. Now that Professor Hinton showed that the problems with machine learning could finally be overcome, you might say the Algorithmic Age had well and truly begun. So now we find AI everywhere and it continues in the future also. From self-driving cars to recommendations for movies and TV, facial recognition, city brains, and augmented reality.
But the type of AI that I’m talking about are not the kinds that run factories or digital trackers for your health. I’m interested in the potential of virtual assistants that are reflections of ourselves. “But what about smart speakers?” you ask. “Don’t we already have Alexa, Google, Cortana, and Siri? “Voice assistance that can tell us the weather, “turn on a light, play a song, or bring a pizza in a jiffy?” All of these AI services are, in my view, a kind of public rather than private AI. The difference is not so much what they do but more so, the why.
With public AI, all your life’s data is just a small part of a vast database used by tech giants as part of their advertising and the retail arms race. Your data is really not yours at all, and in fact, who knows who else might be listening in? Do you really want to live in a kind of a digital Downton Abbey, with tech staff that knows your every secret and sin? Private AI is the start of something new in the future. Imagine an AI designed to learn, but not with the aim of marketing to you.
With private AI, you own the data, and more importantly, also the future AI model that you train. More than an assistant, it’s a kind of a digital doppelganger for your brain. The idea of a virtual assistant is a seductive one. Ironman had J.A.R.V.I.S to help around, while HAL had his own ideas of how things should get done. As we get to know our personal AIs, will we fall in love with them, too? Well, in Japan, you can buy a holographic virtual girlfriend. Or as in the movie “Her”, you can just download Scarlett Johansson.
The AIs I’m talking about is not quite like these. Rather than a distinct personality, imagine an AI that simply reflected you and your needs. A true extension of yourself, designed to handle small decisions or answer your mail, act on your behalf, and can be trusted, without fail. One of the earliest examples of a personal AI might be this one from Deepak Chopra designed to be a spiritual guide. Is this a new form of immortality? Not just a hologram or the singularity but an AI that can speak for us after we’re gone it shows really great future. Maybe, but first, there are a few challenges that must be overcome.
1. How do we ensure the security of our digital twin?
In this new era of cyberattacks, ad tech, and malware, it’s never been more important to understand security and the data we share. What we need in the future to AI built by and for us alone, whether it runs on a home server or directly from your phone. For more complex processing, our personal AIs may require a private cloud, but it shouldn’t matter where the data lives as long as it’s clear who owns it and what uses are allowed.
2. How will AIs get smarter if they can only learn alone?
Some believe that one person’s life and personality is not enough data to train AIs that rely on scale and diversity. Actually, that’s no longer true. When it comes to machine learning, there’s already been a number of fascinating breakthroughs. For example, consider federated learning and differential privacy, new techniques designed to protect your data and identity. The price of an AI getting smart shouldn’t be advertisers being able to pick your life apart which is possible in the future.
3. Can we turn our private AIs into a source of personal advantage?
In the future, our personal AIs will be more than just a way of saving time. They will augment our capabilities in ways still hard to define. (sonic beeping) Will training our AI be part of what we do? Will we even start bringing our AIs to school? And when you go for a job interview, will they ask your AI questions too? Will wealthy families compete with their own finance AIs, like an algorithmic investment fund designed to optimize wealth for the long run? Speaking of money, just how much are the lawyers going to make when people start trying to use AI as a kind of virtual estate? In this new age of smart machines about to dawn, our augmentations may be as important as that with which we’re born. And when we can no longer imagine life without AIs that act upon our whim, will you be able to separate yourself from your own algorithm? With this, we come to the end of How can AI affect personal life in the future?