A real question that opens up a divide between those who are in favour and those who consider that AI could represent a threat.
Before developing an argument on this topical subject that attracts the interest of many players. Let us first define what artificial intelligence, which everyone is talking about without always really understanding how it works, actually is.
TABLE OF CONTENTS
- How can artificial intelligence be defined?
- Where is this artificial intelligence in our daily lives?
- A few historical reminders
- Science or recklessness? Should we be worried?
- AI and failure?
- Which sectors are most likely to be automated?
- Those who fare best
- The opportunities of artificial intelligence
How can artificial intelligence be defined?
Specialists agree that artificial intelligence can take various forms: emotional, organisational, systematic, rational, literary, strategic, creative, situational, etc. So we cannot speak of ONE single intelligence, but of a multitude that take various forms.
It is above all a scientific discipline, concerned with the processing of knowledge and with reasoning, in order to give a machine the same capabilities as a human being in terms of intelligence.
If we go deeper into its definition, we can distinguish four main pillars:
- Self-learning (the ability of a machine or a computer to learn autonomously, over time, one might even say through “experience”)
- Audiovisual recognition (the ability of a machine to have sensory faculties just like a human being, through sight, hearing or touch)
- Natural language processing (the ability of a machine to interact with a human, in speech or in writing, in natural language)
- Task planning (the ability of a machine to draw up plans, to plan series of actions).
Where is this artificial intelligence in our daily lives?
The development of this artificial intelligence has brought out a large number of products that are increasingly an integral part of our daily lives. We could mention personal assistants, which are becoming ever more numerous, more advanced and sharper.
Google is making its mark with the Home assistant, which is on a completely different level from the previous assistants available on Androïd (Google Now, Isis Assistant, Jeannie, Skyvi). Although Isis Assistant draws on the features of Siri for Apple (non-exhaustive), it is far from reaching the same performance. With Home, Google makes it possible to link all the devices in the house, which are now managed by voice and interconnected.
At the same time, Amazon Web Services is launching Rekognition, an image-recognition artificial intelligence able to identify elements such as faces, emotions, gender, the number of person(s) and many other things. The system, having recorded thousands of images to build itself a base of knowledge, can in particular recognise an animal, and even its breed. It is built to carry out processing in real time, or in batch.
In the end, with a little more attention, we realise that artificial intelligence is taking up more and more space. But none of this dates from today…
A few historical reminders
In 1956, the first artificial intelligence programme saw the light of day, along with the term itself. It was only in 1962 that the world witnessed the first victory of a machine over a human in a game of draughts. A little before the 2000s, in 1997 to be precise, Deepblue, a computer developed by IBM, managed to beat Garry Kasparov, the world chess champion. In 2010, Google launched its first self-driving car, which relies in part on artificial intelligence. 2016 marks the year of artificial intelligence, declares Microsoft, and the CEO continues along the same lines, adding “It is our responsibility to make artificial intelligence augment human ingenuity and human possibilities”.
The biggest challenge of tomorrow for researchers is to design emotional intelligence, which could recognise a person who is satisfied or dissatisfied, through voice and morphological analysis. In time, AI could therefore adapt its language to the personality of the person it is talking to, thanks to a library gathering the various expressions that a human face conveys, or by decoding the human voice.
This technology has been constantly renewing itself since its earliest days, so how far can this intelligence go?
Science or recklessness? Should we be worried?
AI is not seen as a threat by some, because, as we have seen, it only performs well in a very specific field and context. This non-existence of ONE single artificial intelligence is enough to reassure a section of opinion, which sees in it an opportunity to improve the performance of their company. Indeed, Forrester reports in its forecasts that investments in AI are set to triple during 2017. Artificial intelligence gives them access to valuable information, out of reach until now, all thanks to cognitive interfaces built into complex systems, particularly in terms of machine Learning.
The main target of these investments is customer information, with a view to becoming “insight-driven”.
AI and failure?
A recent event has nevertheless shown recently that AI can have certain drawbacks. Here we are referring to the AI developed by Microsoft, “Tay”, a “chatbot” that was handed the reins of various social network accounts last March. The experiment was a fiasco, the AI was the subject of a hijacking operation organised by, among others, activists supporting Donald Trump, according to the newspaper Le Monde. The experiment turned into a fiasco through their action, “Tay” made more than inappropriate remarks. The next day, the AI was deactivated until it could be put right.
That is why “65% of French people say they are worried about the development of artificial intelligence” according to the result of an IFOP survey. That concern does not, however, slow down the deployment of AI, though its impact differs from one sector of activity to another.
Which sectors are most likely to be automated?
In first place is food service, with automation that could reach 73% according to McKinsey, followed by: production at 59%, retail and transport at 51% each, and resource extraction at 50%.
As a result, the most exposed jobs are blue-collar and administrative jobs. They are not the only ones. We can also mention knowledge workers active in different sectors, such as finance at 43%, healthcare and the media both at 36%, with writers squarely in the firing line.
Conversely, those who fare best
In terms of feasibility, in the education sector the share of AI should not exceed 27%, and it is lower still for management activities, with an automation potential of only 9%. Decision-making and planning are therefore still tasks that we visibly struggle to automate.
According to the OECD study, 9% of jobs in France could be replaced by robots and/or software. This artificial intelligence trend is due to the growth of the IoT. By 2020, Mckinsey & Co puts the number of connected devices at 30 billion, which represents an immense data opportunity, with the new challenges of collection, processing and action that follow from it. AI could turn out to be an indispensable means of processing this data. Which both confronts and at the same time supports the work of data scientists.
The opportunities of artificial intelligence
AI seems to be a support in carrying out an employee’s tasks, but it can also be a permanent and cheaper substitute. We are very clearly entering the creative destruction set out by Schumpeter, AI is certainly a substitute solution but it is also a creator of jobs. 30% of positions in robotics are said to be unfilled, which offers a fairly bright future to developers, and to all the professions that make it possible to manage this AI.
The public Cloud giant never stops surprising us, for 20 years it has been dedicating thousands of employees to this promising branch. So with its new artificial intelligence services, Amazon Web Services seems to want to reposition itself against its competitors, Microsoft, IBM and Salesforce, which all seem to have had a nose for the added value AI could bring to their Cloud offerings.
Premaccess has become a reference partner for AWS, and together we give you the chance to make a projection into the AWS Cloud.
Do not miss this chance to reach the ultimate cloud, and all without paying a thing.
Contact us!
So let us expect a few new things from the Cloud giants before long.
In the end, should we fear it?
Although the study “Artificial Intelligence and Life in 2030” published by Stanford University last September shows no risk concerning artificial intelligence. That is not the view of the scientist Stephen Hawking and the philosopher Nick Bostrom, or indeed Bill Gates, who have expressed great concern on the subject, even calling it a “danger for humanity”.
In the end, we are a long way from a “Terminator” scenario but only the future will prove it to us…