By Göran Karlsson, Futurist, Executive Partner and Business Developer at TCS
The potential of AI fascinates me. You may not necessarily be aware of it, but AI is working away in the background during the many routine tasks we carry out every day.
The power of AI is gathering and processing information when we’re buying something from Amazon, streaming a TV series or planning a vacation. In a business setting, it can help better understand the needs of customers, or automate production at a factory.
And the reason why AI is so involved in all these experiences is because it is essentially an optimizer that can be leveraged to improve any process, anywhere, making experiences more relevant.
However, the role of AI – and the data it is collecting and crunching – is not transparent to the majority of consumers.
And that has led to calls to make sure that the use of AI is both ethical and accountable. The European Union (EU) is establishing a regulatory framework as part of that drive.
At TCS, we have also started work on a framework for ethical AI. After conferring with stakeholders at Think Digital in November 2018, we’re due to publish our Point of View early in 2019.
It will work as an important basis for further discussions with our European customers, as well as feeding into the wider EU framework.
The EU guidelines talk about the future of work, fairness, safety, security, social inclusion and what is called ‘algorithmic transparency’.
However, although the regulation is undeniably important, we need to make sure that inertia doesn’t set in.
People tend to think that AI is dangerous, just because it’s new. Businesses may also be quite conservative, especially when they hear the discussions between industry and politicians around the governance of AI.
All this means that we could be in danger of being overcautious and could end up hindering its potential. In Europe, we already have some catching-up to do – of the top 50 global tech companies, none of them are from the EU.
However, the evidence from a survey that Tata Consultancy Services (TCS) conducted about AI a few years ago suggests that the EU is not actually falling behind when it comes to innovation in this area.
We need to make sure that the political landscape is actively accelerating the advance of AI rather than just keeping up.
The fuel of data
Data is one area where it has already become apparent that clear guidelines are needed to make sure an ethical approach is being adopted. And the General Data Protection Regulation (GDPR) rules have made a good start in this respect.
But AI, of course, is only as good as the quality of the data available to it.
Simply put, AI works in three distinct stages: collect data, optimize and learn, and ultimately perform business actions. It analyzes and understands vast data sets to find the patterns that will lead to optimization.
A lot of companies have more than sufficient data, but it’s hidden in silos. When these silos are opened up to AI, businesses will become much more effective than when they are basing their decision-making processes on assumptions.
In each and every AI project, data needs to be collected as the fuel for the optimize and learn phase.
This is where ethical issues – and legal implications – come into play.
The collection of data, its classification and verification must be secure and trustworthy. And policies to ensure quality, security and unbiased data are therefore critically important.
Within this cycle of optimizing, learning and re-optimizing, there needs to be a set of parameters or rules to define what the desired results are. At the same time, AI needs to be able to proceed with an acceptable degree of precision if it is to be effective.
The future of work
There is also an ethical dimension when it comes to preparing the workforce to cope with the forthcoming wave of change.
The financial services industry is one of the areas where we’re seeing the most significant uptake of AI. In London, for example, machines have pretty much-replaced humans when it comes to broking deals and speculative trading.
The human brain only has limited capacity, but AI can detect a potential fraud, for example, by examining many differences across multiple dimensions.
Such changes lead to justifiable fears about wider job losses and redundancies. But, in the same way other technologies in the past have enabled growth, the introduction of AI represents a shift towards more meaningful jobs – and it will create new occupations that didn’t exist in the past.
AI lacks emotional intelligence and empathy – both among the uniquely human skills that are irreplaceable. But in order to optimize the potential of AI to improve profitability, we have a duty to retrain and reskill our workforce.
At TCS, we’re investing massively in training our 420,000 associates, using state-of-the-art facilities. The French government’s 2018 Villani Report focuses on four key categories of skills and abilities we need to enhance: Cross-cutting cognitive skills, which includes problem-solving and understanding language and numbers; creative abilities; social and situational skills, such as teamwork and independence; and precision abilities, such as manual dexterity, which should not be overlooked.
So an ethical approach means ensuring people are retrained and adequately reskilled. And the EU has an important role to play in incentivising people to get the appropriate level of expertise in science, technology, engineering and mathematics (STEM) subjects.
In summary, we need regulation but we also need more incentives and education. And – if we are to avoid inertia – we also need to embrace a degree of risk.