Illustration: © IoT For All
The ever-increasing adoption of AI-powered systems in all areas of the economy could either lead to potential discrimination against women or open up new career prospects. Let’s find out what needs to be done to achieve the second scenario.
For years, there have been debates that artificial intelligence will change the labor market. Nowadays the focus has shifted from the inevitability of the future to an assessment of how and when the world will change and what it means for all of us. There have been several attempts to create AI, but Artificial General Intelligence – the one portrayed in films and books – is still a long way from being built. What do we have now?
Not the Smartest Intelligence
There are many narrow ‘smart algorithms’ that solve niche tasks and can find patterns in large amounts of data. They help to make decisions on, for example, whether a suspicious transaction should be stopped or allowed, where cargo should be sent, or whether a loan should be granted. The probability of a mistake in these systems is lower comparing to a human being’s performance. Further development of algorithms capable of supplementing or replacing human decision-making will reduce the number of roles requiring routine decision-making, which are often taken on by women. It is already difficult to imagine a bright future of tour operators, taxi dispatchers, clerks, and some other professions. According to a forecast made by IMF researchers, 11% of jobs held by women will be automated within the next two decades. The focus will also shift into designing these algorithms.
More and more decisions are made automatically and, sadly, not in women’s favor. For example, two years ago Amazon, the largest online retailer in the world, was forced to ‘dismiss’ an automated CV-reviewing system because it discriminated against female developers. This situation occurred because the data used to train machine learning models represented a hiring history for 10 years. At that time, the majority of candidates were male, but this does not prove that men are better at technical work. Considering that algorithms are becoming in charge of making an ever-increasing number of decisions, similar side effects can be an issue.
It is widely believed that algorithms just find patterns within an existing data set. However, focusing on data, it is easy to forget two aspects of this problem: the limitations of existing algorithms and, more importantly, the role of people who train them. Most algorithms just catch the correlation within the data, without understanding anything about it. Even the best data is meaningless as long as people who can solve problems and ask the right questions are not involved, so the algorithms will simply reflect our own biases.
Developers of AI systems must carefully monitor the way datasets are formed and track any biases that might occur. Mistakes made by the algorithm should be tracked: sometimes the percentage of errors is quite low, but they could be related to one group of people. For example, a scoring model systematically refuses to give loans to residents of Chinatown. Such behavior is particularly dangerous because we are shifting more and more responsibility for decision-making onto the system. Some advanced algorithms which are currently in use cannot even be interpreted; i.e. we are unable to understand why a certain decision has been made at all, or which factors influenced it.
Set The Rules
It’s interesting to note that the rise of computer science is strongly connected with women: Ada Lovelace created programming, Betty Holberton designed the first general-purpose computer, and Margaret Hamilton developed the software for the Apollo project. Today, only 15% of specialists in the field of artificial intelligence around the world are women, which is disappointing for many advocates of gender equality. Some people worry that the future will be created by men for men. The presence of female analysts eases the situation since they can spot the problems with products that are not easy to discover if you don’t face discrimination on a daily basis.
Women already spend a lot of energy ensuring that they are treated honestly and fairly. Perhaps a legal base is needed. Some regulatory bodies have already taken up this issue. For example, the EU General Data Protection Regulation (GDPR) requires companies to explain the reasons behind decisions made by AI systems and to monitor them to prevent any discrimination. Currently, only large companies and governments can afford the software and hardware required to launch the most advanced AI models. This barrier could be used to help set up some basic rules before the technology becomes widely used.
New Skills; New Challenges
Many futurists are optimistic about women’s chances of success. Why? The most valuable skills in this brave new world are those that algorithms cannot master: the use of soft skills and emotional intelligence. As such, 83% of organizations surveyed by Capgemini believe that emotional intelligence will be a prerequisite for success in the coming years. The new labor market will value compassion, multi-tasking, cooperation, and empathy – traits that are traditionally associated with women, meaning that women will have greater chances of being hired. However, there is a nuance here. To succeed in an AI-ruled world, a great work of retraining and adaptation is necessary. After all, it will be not the strongest and intelligent that survives, it is the one that is most adaptable to change.
Extra risks for women will be caused by an inability to adapt. The new challenges of automation are added to conventional difficulties, creating barriers to gender equality. Nowadays there is a need for mobility and flexibility among employees as it is now easier to change profession, employer, industry, and even country than ever before. Women are often less mobile than men, because of their ‘second job’ at home. What’s more, they are often excluded from networking, which allows men to improve their skills, find mentors, and new employment opportunities.
On the other hand, companies will be more motivated to push their employees to develop skills that cannot be automated. If women are proactive and able to adapt, they will have more employment opportunities. Bear in mind that being good at something no one needs is the biggest waste of time. In the future, two categories of skills will be most important: the ability to negotiate and standing your ground, and the ability to see trends and build strategies.
Organizations need employees who can talk to both machines and people. Recently, technologies have been significantly democratised, meaning that you do not need a PhD to work with AI. Today is the best time to gain insight on the foundations of a technology which at first seems complicated, even if your profession is not connected with data analysis at all.
It would be useful for executives to at least make themselves comfortable with the methods of machine learning for analyzing different types of data. Such analysis will make it possible to assess use cases for AI implementation in a specific area and to build an effective strategy for digital transformation. A deep dive into the topic and acquisition of practical skills, such as programming and creating machine learning models, will be useful for those who spend a lot of time running routine analysis and who want to automate the decision-making process.
Implemented projects from industries that are more mature when it comes to AI adoption, such as IT companies, banks, and retailers, can be good sources of inspiration. Even such a conservative industry as manufacturing has started digital transformation programs for production optimization, for example, to forecast machinery breakdowns.
According to IDC forecasts, by 2021 AI systems will be implemented in one form or another at 75% of enterprises. So, the relevant skills will be required in almost all industries. Allocate an hour and watch a video on the principles of machine learning. Only by having a clear understanding of how it works, it is possible to ask the right questions and set goals. Knowledge of a few principles is more important than understanding detailed implementation ways for all algorithms in a software package.
Once you understand the basic principles, it’s important to understand which tasks are the most relevant for applying AI and to try delegating some responsibilities to the machine to increase the quality of work and performance. By studying successful scenarios of AI application, many people realize that this technology can be used for a wide range of processes and tasks that involve working with large amounts of data. Women should learn how to find prospective use cases for new technologies, have sufficient motivation to look for answers, and take under control their education and career.
Data analysis skills are relevant to a huge number of professions, from marketing experts to mechanics working with a CNC machine. For example, AI can identify anomalies in a technological process. Currently, there is a lack of personnel in data science and artificial intelligence space. The problem is so urgent that large companies are offering free platforms for studying. After the training, you may not be able to perform as good as an experienced data scientist, but you will be able to translate a task from a business domain for the researcher, for example, regarding an analysis of CV-collection or even writing music.
According to the Institute of Electrical and Electronics Engineers, an international non-profit association of specialists in the field of computer science, “The need for specialists in artificial intelligence has revealed itself in almost every area of life.” Experts are urging for AI training for specialists in areas such as healthcare, agriculture, and logistics.
Conventional methods of working are rapidly becoming obsolete. Humanity has to choose once again: to lament about fate and talk about the rise of the machines or to get ready for the future and acquire the skills that are in demand. The Luddites have lost in their fight against the machines, simply because machines are economically efficient. And, as we know now, after all the number of jobs created thanks to the adoption of machines has been far greater than the number of jobs gone.
The most insightful employees were not afraid of industrialization, and made use of and reaped the benefits of new technologies. So, why can’t women do the same?