Why we are underestimating the impact of automation and AI16 Oct 2017
Azeem Azhar, product entrepreneur and curator of weekly tech newsletter, Exponential View, discusses how AI can deliver real value for businesses and considers the risks related to the emerging technology
AI is a very odd nomenclature. It has a competing technical definition even amongst its real practitioners (mostly in labs). We can agree, however, that it includes some tools such as machine learning, natural language processing, and a product or business quality. In short, it’s a system that learns from experience, or a system that uses any of the aforementioned tools.
Two famous examples of firms using AI for profit are Google and Facebook. Take money machine Facebook: its ad business is entirely driven by the application of AI technologies and machine learning. The tech is employed to help create personalized newsfeeds to increase the amount of time that users spend on the platform. AI is also used to place more tailored adverts in front of those captive users.
Early adopters like Facebook are typically larger, digitally mature organizations which, more commonly than not, have already led in previous areas of digital transformation. They are firms that are highly focused on development over cost savings and their C-level executives support the implementation of AI tools – AI is often at the very core of their business.
But you don’t have to be Facebook to benefit from AI or machine learning. Accenture research indicates that AI will increase profitability by 38% across multiple industries by 2035.
In firms that want to catch up on the wave, executives should start scanning use cases and analyze what models need to be introduced to suit their own business goals.
Digital industries will benefit first with the technology allowing the IT sector, for example, to advance through the automation of processes and more effective prevention of cyber attacks.
Manufacturing will also see improved productivity through the use of robots and co-bots (collaborative robots), and financial services will benefit from better risk management and automation.
Startups is another key area where AI innovation can flourish. There are some very interesting AI products being delivered by startups across a diverse range of domains, including factory automation, pregnancy management, and traffic routing.
Every sector should be able to benefit from AI technology, even those which appear to be less natively digital.Risks to employment
The transition to an AI-first world will certainly be painful for a large number of people. In the short term, millions of jobs will change.
Occupations that deal with repetitive and predictable tasks are going to be dealt the first blow. Retail will certainly feel the hit with the growth of e-commerce and the introduction of intelligent systems which help to process orders and deliveries. Data collection and processing are high on the list too.
On the other hand, jobs that require managerial skills, empathy, expertise and adeptness to ever-changing circumstances will not be automated for a very long time. Neither will jobs which require a human touch, such as beauty therapy.Timeframes matter
The capability of AI systems is increasing rapidly and so too are the number of tasks that they can complete. Take, for example, compliance in investment banks. A large part of this role is being able to listen into phone calls to identify illegal or problematic dialogue. Today, these communications can be analyzed by speech-to-text technologies in near real-time. Suspicious remarks can be flagged immediately. This is a huge area of employment now at risk from a technology which simply did not exist five years ago.
Businesses must now consider how quickly automated systems could ascend into their domain and how prepared they are for adopting AI at the expense of humans.
We are likely to underestimate the impact of automation on the workforce in Western markets over the next ten years. It’s a hunch but if it’s right we probably need to be acting sooner rather than later.