The history of artificial intelligence dates back to ancient Greece. Philosophers of that time were already wondering whether human thinking could be performed by intelligent machines. This question has been with mankind for centuries. The answer to it came in 1955, when Herbert Simon and Allen Newell, American scientists, finally created the very first artificial intelligence program.
Today, over sixty-five years after this breakthrough, we have developed algorithms and machines that our ancestors could only dream of. But, even now, we haven’t fully mastered AI technology. There are still many areas of AI that people need to understand better. Fortunately, research into artificial intelligence has accelerated recently.
Over 50% of respondents who took part in PwC’s 2021 AI Predictions Report say that their companies increased investments in AI over the last year. Businesses invest in AI and machine learning with the hope of improving their customer satisfaction and productivity, and to decrease the risk of bad investments. Recent developments have only strengthened already emerging trends while also creating new ones that promise to have a great impact on the near future.
2021 AI and machine learning trends
1. AI democratization
Up until now, artificial intelligence has only been available to tech giants and startups, and they have left the rest of the world far behind. That’s going to change. According to Gartner, the democratization of AI will be a dominating trend in 2021. That means that more organizations will put more effort into making AI tools accessible.
The more popular AI tools will become, the more data they will collect. That will help to make AI products more accurate. What’s more, AI will become cheaper, and because of that, it will become obtainable even in sectors that haven’t been able to afford using it so far. Organizations benefiting the public, state schools, and healthcare facilities could fully benefit from it and also get a chance to become much more effective in their operations.
However, before that can happen, businesses must act responsibly. They need to put extra effort into training employees at all levels, not only the technical staff, so that everybody can benefit from using the technology. This will help to reduce the effects of AI illiteracy that could become a huge problem in the labor market of the future.
2. AI industrialization
AI democratization won’t make it without AI industralization, which is the trend of making AI products reusable, scalable, and, simply speaking, available for the masses. According to Pega’s report, the Future of Work, 82% of frontline IT staff believe that more companies developing AI will share their technology as low-code products. AI pioneers that have been developing AI technology for their own benefit, will become educators. They’ll be helping smaller players make use of already existing AI tools. In effect, fewer organizations will need to build AI solutions from scratch, and that will speed up AI adoption in the market.
The pandemic highlighted the need for workers’ safety. That sped up the robotization trend that has already been visible in many sectors. Machines can perform many tasks for humans like helping businesses and institutions improve efficiency and to help maintain social distancing efforts. Take, for example, Walmart and Amazon. The companies have been using robots to sort and pack products and to clean spaces. Since the pandemic started, Chinese hospitals have been using robots to measure the temperature of patients entering the complex and to disinfect the buildings.
In Singapore, the police started using Spot, the robot dog created by Boston Dynamics, to monitor the traffic volume on the streets and remind pedestrians about pandemic restrictions.
The use of machines in various sectors will speed up the manufacturing processes, reduce the number of workers who need to physically turn up at work, and let businesses avoid downtime during lockdowns.
4. Occupation transformation
There’s no doubt that technology improves human productivity. However, in many cases, it can take over a job performed by a human completely. McKinsey’s report predicts that by 2030, there will be up to 375 million people who will switch their occupation due to the adoption of automation. Should we be afraid that robots will steal our jobs? Not necessarily.
While many jobs will be lost, others will be created. The researchers estimate that by the end of 2030, there will be up to 365 million new jobs created because of growing incomes and consumption. People will become richer and will consume more goods and services. This will create new jobs in emerging markets and countries that manufacture and deliver them. The industrialization of AI will also create a demand for specialists who can implement AI technology and educate people on how to use it.
5. Responsible AI
The increased use of AI and machine learning will make businesses and governments pay more attention to their ethics. Over the past 10 years, we have observed many cases of AI bias and discrimination. For example, Compas is an AI system that has been used by the American court system to forecast which defendants could reoffend. The algorithm was predicting that black offenders were more likely to become recidivists than white defendants. That wasn’t true. Because of the data Compas was initially provided with, the algorithm developed an incorrect model and was misclassifying black defendants.
Another example that helped show the limitations of AI was Amazon’s recruiting tool. In 2014, the company started testing an AI system that was meant to help their recruiting team search for top talent. The problem was that the algorithm was learning by using the number of applications submitted to Amazon over the past decade, and these belonged mostly to men. Because of that, the program taught itself to prefer male applicants, penalizing CVs sent by women.
These are just two examples that show that making unbiased algorithms is still challenging. More and more employees, customers, and investors are now expecting business to address the ethical issues surronding AI and to make the technology explainable and transparent. Companies will also have to decide to what extent they can hand their responsibilities over to algorithms. Should those algorithms be able to make choices for humans, or should they only provide advice? If a mistake happens, who should be responsible for it? This pressure will, sooner or later, make governments create regulations that will ensure the fairness of AI and its accountability.
‘’In 2021, there will be an increasing hue and cry about the continued unearthing of AI For Bad, and a realization that the only path to AI For Good involves turning the “talk” about AI Ethics into the “walk” for AI Ethics.’’
Dr. Lance Eliot, author of the AI Trends Insider on Autonomy column in AI Trends.
6. AI in Healthcare
In 2021, AI and machine learning will also be put into service in the healthcare industry. Hospitals have already been using it to plan their budgets and improve treatment efficiency. But, they’ve been doing more than that, too. In 2020, scientists from Harvard and MIT used machine learning to measure the impact of the pandemic on peoples’ mental health. The algorithms analyzed over 800,000 Reddit posts to track whether there had been any changes in users’ tone and language. The results showed that the number of threads about suicide and anxiety more than doubled during the lockdown when compared to the same period of the preceding year.
The researchers say that, in the future, similar analyses could help psychiatrists learn more about different mental illnesses and ways people demonstrate them. What’s more, forum moderators could use it to identify users who might urgently need help. By doing so, they could connect with these users and contact them with specialists that can help them.
Machine learning and artificial intelligence have proven to be useful in drug discovery, too. A biotech company, Deep Genomics, has already used it to discover a genetic mutation that causes the Wilson disease. It’s a serious disorder that causes the accumulation of toxic levels of copper in the body. The Deep Genomics’ discovery will help to create better therapies and shows that AI has a huge role to play in medicine.
The wider use of AI will bring new opportunities. At the same time, it will make hackers and cybercriminals more efficient. Recent cyberattacks have shown that cybercrime can be as difficult to stop as the pandemic itself. In 2020, in the U.S., the Department of Health and Human Services and the Office for Civil Rights detected 436 data breaches that affected over 17 million people in the American healthcare sector alone. Medical facilities are especially vulnerable to hacker attacks because the sensitive data they store sells well on the dark web.
On top of that, there are more and more tools that facilitate cyberattacks and make them easy to commit, even by non-advanced coders. Because of that, organizations will have to use additional resources for cyber programs in order to prevent attacks. According to PwC’s study, over 55% of enterprise executives are going to increase their cybersecurity budgets in 2021 and hire new cybersecurity staff. By using pattern recognition to monitor their websites, businesses will be more likely to mitigate the risk of attacks.
8. Chatbots and voice assistants
Lockdowns have made many companies go online, and that increases how much they spend on customer service. This has accelerated the adoption of chatbots. Data shows that just in March 2020, the weekly use of chatbots rose 426% compared to January 2020. The trends show that customers want to contact brands with their mobile phones and when they’re on the go. As chatbots integrate with various communication platforms, they can help businesses meet these expectations.
A similar trend applies to voice assistants whose growing popularity is driven mostly by younger users. The 2020 PwC’s report shows that younger generations prefer hands-free conversations, and that’s why they are more likely to experiment with voice technology, as opposed to older users. The majority of consumers are eager to use voice assistants to handle simple tasks like ordering pizza. However, when it comes to serious shopping or making a complaint, consumers still prefer to do it via channels they already know and trust.
Some people say that artificial intelligence is changing our lives as significantly as the ways the electric lightbulb did. Once we get used to using it, we can’t imagine living without it. There’s nothing wrong with that. The important thing is that now the people responsible for creating AI technology should do their best to make it transparent, ethical, explainable, and accessible. Only in this way will societies equally benefit from artificial intelligence and narrow the gap between those who grow thanks to AI and those who are left far behind.
Hungry for more data?
Check out our 2021 Chatbot Statistics and learn even more about the trends shaping the future of technology.