iKala Future Talks

The Latest Innovations in Artificial Intelligence – Part 1

Play Video about 趨勢相談所

According to IDC, the artificial intelligence market is expected to break the $500 billion mark by 2024, with a five-year CAGR of 17.5%. What are the top AI innovations in 2021? When AI is revolutionizing and transforming human society, how can we respond to the challenge and benefit from AI? 

Here are some highlights from iKala Co-founder & CEO Sega Cheng's sharing.

Challenges of Adopting AI

From our previous experience, one of the pain points of people adopting AI is that they spend more than half of the time processing data. In the first half of the project, people should spend a lot of time and effort on cleansing, labeling or even augmentation of data; while AI models and algorithms only account for a small proportion of time. We find that there are still entry barriers for many enterprises to adopt AI. 

AI came out around 1960s. As you can see from this curve, the time and cost we spend in training AI systems have been increasing. You can find that after the boost of deep learning in 2010, the whole curve is reversed. We once doubled our computing power to train better AI systems every two years, but it turned out to double every three to four months after 2010. That is why "cost" has been a major concern for the industry when adopting AI and making further breakthroughs on it.

Key Developments in AI

Stanford University  released the "2021 AI Index Report" that excerpted highlights of AI's development around the world.

  1. AI investment in 2020 is 4.5 times higher than 2019, probably because of COVID-19, people invest in AI research focusing on pharmaceutical and biomedicine. Investment on drugs, drug discovery and molecular research in 2020 sees a massive increase.
  2. From 2019, 65% of graduating PhDs in AI went into the industry, compared to 44% a decade ago, when deep learning just began to emerge.
  3. AI has a certain creative ability that can compose text, audio, and images.
  4. AI is now facing a diversity challenge in researchers and industry workers. In the United States, 45% of resident AI PhD graduates were white.
  5. The AI industry in China has been developing rapidly in recent years due to the large amount of data from its government. China has surpassed the United States in the total number of journal publications for the first time in 2020. However, the United States has consistently more AI journal citations and conference papers than China over the last decade. We can tell that China and the United States are firing on all cylinders for AI.
  6. In the United States, international students in the AI field rise to 64.3%, and this proportion continues to increase.
  7. Surveillance technologies have evolved quickly in the past two years, mainly because the accuracy of computer vision has exceeded that of human eyes in 2017. Governments are also adopting different kinds of surveillance technologies, and this raises some ethics problems and social issues.
  8. As AI is under rapid development, the discussion of public policies and personal privacy just began in the past few years, so that AI ethics lacks benchmarks and consensus.
  9. AI finally gained the attention of the US Congress. The 116th Congress is the most AI-focused congressional session in history, paying great attention to the international development of the US internet giant companies. In addition, many congressional hearings related to AI and digital technology were raised.

AI's Impact on Jobs

Looking back to the last two decades, some scholars began to investigate on how susceptible are jobs to computerization as computers and artificial intelligence become more and more mature in 2013. The left side in this picture stands for relatively unaffected work and the right for high-risk ones. We can see that those on the left are jobs related to management, computer skills, education, and health care, which require some soft skills so that it would be hard for them to be replaced by computers. However, for these red areas on the right, it may be jobs of manufacturing, transportation, or even agriculture, forestry and fisheries, which machines can easily involve in their work, and they are classified as high-risk jobs.

On the left of this chart, we found that like telemarketers may be at high risk of being replaced by machines, but if you look upwards, like the entertainment industry, even clergy and religious personnel will probably never be replaced. Therefore, automation plays different roles in different job functions. AI can almost do what humans can do to a certain extent, so the key point is whether the work is repetitive, and the automation brought about by repetition.

Besides replacing jobs, AI also creates new jobs. A PwC study in 2018 indicates that in the UK over the next two decades, in the area of health care, there will be more jobs created than replaced. 

A Double-Disruption Scenario for Workers

The World Economic Forum actually released a report last year. The impact of the epidemic is actually a double blow. Business owners will start to consider reducing manpower, and consider office automation to further cut costs and survive the pandemic. The overall environment is ever changing due to the pandemic, and there is a high degree of uncertainty, so at this time, 41.8% of business owners will begin to increase the use of contractors.

Vital Skills in the Future

When Harvard Business Review asked managers what the most important skills in the future are, most of them answered "digital technology", but in fact, the real situation is that social skills, the ability to interact with people, guide, teach, and cooperate with each other is underestimated.

Even in the field of mathematics, for mathematicians with social skills and those without social skills, there's a gap of up to 15% between the average salary they get. This is an example of why we should pay more attention to social skills in the AI era.

Sign up to watch the latest episodes of iKala Future Talks for free.

Leave a Reply

Your email address will not be published. Required fields are marked *