Artificial intelligence, AI, is the next big technology to have entered mainstream consciousness. From eerie androids such as Sophia to the silent efficiency of automated delivery systems in modern Amazon warehouses, the growth of autonomous driving and the popularity of smart speaker systems such as Alexa or Google Home – AI is everywhere.
And it’s coming for our jobs, white collar and blue, threatening massive social and economic upheaval.
But what is AI really? Why has it suddenly become so popular? Why is everyone so excited about its tremendous potential? Will it really replace humans – and should we welcome it with open arms, or fear for its impact?
Far from being an omnipotent, autonomous robot, AI is at heart simply a machine programmed to make sense of data on a scale humans can’t deal with. It is the king of the algorithm, a machine learning from its own experiences, objective-oriented and highly intelligent, producing logical conclusions based on input. As part of the digital technology connecting people, things and machines on a big data platform, it has the potential to enable solutions saving time, energy and lives, opening up opportunities as yet undreamt of. And it is still in its infancy in its real world deployment.
The use of AI is growing dramatically right now in response to extraordinary increases in the amount of data produced daily, as powerful computing has become available at lower costs. Humans alone simply cannot process the complexity and ongoing volume of data from people, devices, sensors and machines. In parallel, there is a growing awareness of the tremendous potential of AI technologies to solve problems across all industry sectors and the entire spectrum of human life.
AI can unlock scale and opportunity to deal with the grand challenges facing the world today, from ageing populations to sustainable urban living, access to food, healthcare, water and education, reducing poverty and increasing gender equality. Physical AI will be able to free humans from mundane, routine tasks, allowing them to concentrate on more important, higher-end work, releasing creative potential.
In emerging markets and smart cities alike, AI can help overcome natural limitations to growth such as geographic size or lack of natural resources, creating new markets and new value, rather than merely improving on existing models.
Improvements on current models will, however, be where the power of AI is first felt, in its promise of enormous cost savings, increased productivity, lower production cycles and improved back end or internal processes. Within the telco industry itself, AI will accelerate the evolution of network operator infrastructure into intelligent networks able to offer smarter, faster and more scalable services. Using the engine of big data, AI will enable multiple, diverse and often sector-specific demands to be met through highly-tailored network slices managed in real time.
In the financial services sector, for example, AI can reduce the hundreds of thousands of hours needed to carry out regulatory compliance to a matter of seconds; or the time, effort and investment necessary for a mortgage to a few minutes. New financial services may include mass market personalised services, opening an enormous market of lower earners, or microfinancing for the unbanked. In call centres across a range of industries, AI can work either alongside humans analysing complex data sets in parallel to the human customer-facing contact, or take calls as a co-worker as far as possible before passing on to human expertise.
In all cases, AI is a tool to augment human abilities rather than replace them. And it is only as good as the person inputting information and parameters into its system.
This is one of the principal challenges: ensuring that AI is provided with initial information in a way that does not reflect and perpetuate inherent bias, unconscious or not. It is critical to be aware of, and work to avoid, replication of existing divides and inequalities: on gender, race, geography, the urban/rural split, access to education, investment in infrastructure, the availability of talent, the provision of adequate cyber security. Without action, AI will prolong or deepen these divides. There is a very real danger that the powerful impact of algorithms actuated by AI will remain limited to the developed world due to a lack of infrastructure, advanced networks, open data or data scientists.
Providing open public data and open APIs to allow private companies and individual developers to create solutions for public and commercial services is key to democratising AI – and fast-tracking its deployment. Accessing large data sets in the ecosystem to improve quality of life must be balanced against data protection, privacy and security issues.
Preparation in general – and education – is critical. The international community, government, businesses and individuals should be as ready as possible for the seismic changes that the widespread adoption and deployment of AI will bring with it.
The big one, of course, is the transformation of the existing labour market. It is estimated that up to 75% of all jobs will be impacted by AI over the next ten years – and these will not just be routine, low-skilled jobs, but also traditional blue collar sectors such as journalism, law or financial services. Productivity and revenue should rise as costs are cut, but the societal disruption will be enormous.
AI is often invisible, raising issues of transparency and accountability. It is itself a neutral tool, without morality, but the ethics of its use are complex. Establishing codes of conduct and social norms as the first step to any regulation is urgently necessary at intergovernmental, international level. Regulation – as well as the standardisation necessary for it to function in a multi-vendor ecosystem environment – is further complicated by AI’s inherent structure as an active machine, learning in real time with real data.
AI is here – and growing fast. There is an increasingly urgent need to bring together key stakeholders from government, industry and academia to debate its impact on a neutral platform such as ITU Telecom World 2018, the leading tech event organised by ITU, the UN lead agency for ICTs. Making AI democratic, fair and equitable is a challenge that cannot be met by any one single stakeholder.
Experts at ITU Telecom World 2017 last year felt that its first use cases and greatest impact would be economic rather than social: AI will go where the money is, or can be made. In some sectors, if you are not yet using AI, you are two years behind the curve. But the size of the opportunity is so great, the potential so huge, that it is far from too late.
The potential negative effects of AI include social and economic disruption, in particular in the job market; the deepening of inequality; the danger of inherent bias; major issues of transparency, security and accountability; the lack of an internationally-agreed ethical code. Now is the time for contingency plans, for preparation and education throughout governments, industries and societies.
There is downside, after all, to both deploying AI and not deploying it.
AI will be a key component of discussions at ITU Telecom World 2018 in Durban, South Africa, 10 -13 September, providing the diverse perspectives of international experts from government, industry, SMEs and academia.