O’Reilly recently hosted the AI conference in London and whilst attending the general consensus from everyone I spoke and listened to was that no one seemed to be particularly scared or apprehensive about the future.
Kristian Hammond of Narrative Science said that right now is the best time to be in computer science, as computers are everything and soon it will be Artificial Intelligence (AI) that is everywhere as it is poised to send out a message. “That message is that AI is going to change everything we do, disrupt jobs and everyones lives. However, as we move into this era, we are actually not moving fast enough and this is because AI is far too complex for some people to understand.”
Hammond added: “I believe AI will win but that doesn't mean that everyone should be racing towards implementation.” He explained to the audience that we need to take a step backwards and not just focus on the technology, but focus on the functionality. “The functionality is driven by the need and is fundamental to success.”
He finished by saying: “We need to finally make sure the task you are moving forward with is the task you want to complete.” He shared that creators and professionals wanting to work in AI need to ask themselves ‘are you genuinely driven by data, and do you have that data?’ And most importantly he said to make sure you ask yourself if you really do need AI to complete the problem. “Think of it from a human perspective.”
Jason Knight of Intel AI gave a really interesting talk on deep learning at scale and expressed that when it comes down to AI against human nature, that the technology is always going to be more powerful. “Artificial humans are not limited to replicating real humans.”
He explained that now with AI technology, it is very easy to forge and replicate videos that essentially portray ‘fake news’. He then went on to morph a video of Barack Obama giving a speech and overlaid other people’s speeches such as Neil Patrick Harris, Hillary Clinton and Donald Trump to make it look like that was what Barack was saying.
It was interesting to see this as it really did open your eyes to how easy context can be changed with the power of this technology, which led me to think, are we less trusting as technology grows?
Knight believes that artificial humans and AI have the power to transform many industries in the future - especially with the right safe guards that he said can really help benefit, is true as long as you can trust the technology, then surely it help us excel.
I think another key message from O’Reilly’s day was that as with any good duo, it is the power of balance. We need the technology but we also need the human element. Hammond said: “The new technologies are powerful but we need the architects, engineers and technologists to go with them for the best possible outcome.”
One of the most amazing things about AI is that it allows transformation and change.
Data is all around us, it is the way most things now function, but the situation most companies are finding themselves in is what to do with all that data once it is captured. The important issue here is not how you get the data, it is what you do with it.
With data ‘digital transformation’ is the big buzzword that everyone wants to talk about and is constantly thrown around, but why?
Phillip Carnelley of IDC said: “Everyone is worried about some giant company coming along and stealing their business, but this isn't always the case, for every Apple and Netflix out there, there are smaller companies breaking through.”
AI helps businesses deal with large amounts of data, and it is believed that by 2020 digital transformation and data spending is expected to reach €272bn in Europe and this will be based largely on AI.
Machine Learning isn’t new, it’s just a new way to measure data, which essentially is just a special type of code. David Barber of UCL said: “Machine Learning is data driven AI and data science is basically jet statistics for big data, but it is important to remember that machine learning is not the same as deep learning.”
Barber gave a speech on the Thursday morning all about AI and the challenges that remain with it. He said that machine learning is here to stay, and the hard truth is AI is very complex and messy, therefore we need to be cautious with some of our approaches to it.
The trouble with AI is that there can be a lot of fakes. Barber added: “We do need to be aware and cautious of the ethical implications of AI, but sometimes this can just prove too difficult. As humans we can really help and improve this, by highlighting the transformable capabilities of it.”
The message throughout the whole day suggested that AI isn’t about the quickest method and rushing to get to the end, it is about getting the information that shows us how the world works. Of course there are legal and ethical issues about how AI can portray fake news, but it is then up to us humans to determine what is fake and what is real, and this proves that we can work alongside AI and machine learning and not worry about what the future has to hold.
I will leave you with a quote that stuck with me from Barber: “Yes we have made huge progress in perceptional AI, but we are by no means ‘there yet’. We are not done yet and we still have limited progress in general AI, the future is an exciting space especially for AI.”