Ai is already a huge part of our lives
Artificial Intelligence (AI) contributes to 80% of the digital interactions that Americans have online every day. AI is often working in the background of a site like Facebook, Spotify or Netflix, without users ever realizing its importance to the experience they’re having.
Current implementations of AI are being driven by smartphones, among other things. The modern smartphone is essentially a collection of sensors and monitoring equipment. In other contexts, we might call them a collection of surveillance bugs. They record every activity and transaction we undertake in the digital world along with data relating to the current circumstances, from individuals’ altitude, to speed of movement and even, in some cases, what is being said at the time. In the news this week, Google was accused of using our devices’ mobile data to transmit this information back to the mothership. They snoop and we pay, through our phone plans, for the privilege.
The data we create is already used to track and optimize our money, check our credit scores and will be increasingly used in Health Care. In the UK, for example, the government announced this week that they will make public one of their biggest and most unique resources, anonymized public health records. Almost every UK citizen has used the National Health Service (NHS) at some point, many, for the whole of their lives. The details of medical involvements and the results have been recorded. Now they will be studied and used to predict those, in the first instance, who might fall foul of particular types of cancer. The UK government hopes to save 30,000 lives by 2030.
Here comes the next generation of AI
Despite its many advantages, however, existing AI implementations could be described, in simplistic terms, as effective pattern matching. The more data they have to discern the pattern, the better – which is why the data contributed by the worlds billions of smartphones, and which is provided by the NH is so important to worthwhile AI executions.
Judea Pearl is the mathematician who developed much of the conceptual thinking behind the AI which is taught in universities, and used by techno giants to drive experiences on digital properties. For his pioneering work in the 2908s, he was awarded both the Turing Award, the most highly valued accolade in Computer Science.
Since then, he’s been thinking. Now 81 years old, he is proposing yet another new generation of artificial intelligence, one which doesn’t just predict based on similar, disparate events of a similar nature in the past, but one which can discern a cause and its effect.
Pearl believes that providing a machine with insight in to not just what is likely to happen next but why that particular thing is likely to happen will be what gives them ‘real’ intelligence. It’s the difference between knowing that when it rains, it’s more likely that we will experience lightning, and knowing that lightning is caused by particle interactions within clouds.
This is no mean feat – but Pearl believes the most rudimentary aspects of it can be coded within the next 10 years. It will involve, first, digitizing the concepts that make up our reality – here, clouds, particles, lightning, rain, convection and so on. After that, a more conceptually difficult step will be required during which computers will guess the next step of a future state, based on the data they have. The result will be a new level of Artificial Intelligence – where computers really can think for themselves.
Bringing it all together
AI even as it stands, is likely to drive an era of unprecedented economic prosperity, perhaps enough to pay back the debts that countries have sustained in bailing out the banks. These benefits are largely due to Pearl which makes his predictions on the next step in the journey, steps which will largely take place after his own death, credible. His is perhaps the most interesting idea about technology of current times.
The next generation of AI will present new economic opportunities, for sure, but it will also challenge us morally and from a legislative perspective. When a computer can ask ‘what if’ and weigh up many points of view in it’s considered response, it is dangerously close to a legal entity.
When an algorithm can weigh the pros and cons of a decision, it could be reasonably said to have free will, a key characteristic of sentient intelligence. If facial recognition and Facebook leaks to Cambridge Analytica offer us moral quandaries we find hard to chew now, just wait until the son (or daughter) or today’s AI algorithms are birthed.