AI Overview
Esseen tyyppi: Yksilöessee / 2 esseepistettä.
KIRJALÄHTEET
| KIRJA | KIRJAILIJA |
|---|---|
|
AI Superpowers |
Kai-fu Lee |
Introduction
In the last few years, the speed of artificial intelligence (AI) technology development has been amazing. AI has become the hottest topic in science and technology fields. When AI frequently appears in the media or applications, do people really know what AI is? What is the status of AI technology, what will be the impact of AI on the future of mankind, and how people can use AI to create value? AI involves a very wide range of knowledge. This essay will focus on exploring the basic issues mentioned above, letting people get some knowledge about artificial intelligence (AI).
What is AI?
Speaking of artificial intelligence (AI), what do you think of? The movie of I, Robot, The Terminator? Or AlphaGo? Before AlphaGo and Ke Jie’s remarkable Go match in 2017, Kai-Fu Lee predicted that the human win rate is almost zero. In his book AI Superpowers, he described AlphaGo as one of the world’s most intelligent machines. Ke, the best Go player in the world showed his frustration after lost to AlphaGo. Is AlphaGo the AI that one day might surpass the human being in general tasks?
Artificial intelligence (AI) is defined as the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages in the Oxford Dictionary. Artificial intelligence is a typical interdisciplinary subject, involving philosophy, mathematics, computers, cybernetics, neurology, economics, and linguistics. In 1980, an American philosopher John Searle named strong artificial intelligence (Strong AI) and weak artificial intelligence (Weak AI). Strong AI which is also called Artificial General Intelligence (AGI) is aiming to make computers have the same level of thinking ability as human beings. Weak AI also called narrow AI which is a type of artificial intelligence that is limited to a specific or narrow area. Weak AI simulates human cognition. For instance, the Terminator in the movie has the self-awareness and self-consciousness can think as a human can be seen as AGI, and AlphaGo who is only functioning in learning and playing Go belongs to the weak AI. A similar example is OpenAI Five which has been trained for the video game Dota 2 by playing over 10,000 years of games against itself. At present, the mainstream of academic research and commercial applications is the weak AI.
“Machine Learning (ML) is a subset of artificial intelligence (AI). It allows the machines to learn and make predictions based on its experience (data)“. It is one of the many methods that can be used to implement AI. For example, traditionally, to tell a machine that a person’s music preference order is classic music, jazz, and pop music, it’s needed to code these in the program to judge if the music is this person’s favorite. In machine learning, training data will teach machines to find out the method of judging rules. While the user chose classic music, jazz, and pop music 90 times out of 100 times purchase, it found that these are the user’s top selections, the opposite is not. The larger volume of training data, the more accurate the machine drawn the outcomes.
“Deep learning is part of a broader family of machine learning methods based on artificial neural networks with representation learning. ” It is inspired by the functionality of the brain cells called Artificial Neural Network (ANN). Deep Learning (DL) is good at solving complex problems such as image classification, natural language processing, and speech recognition. One common application is the virtual assistant Siri.
What AI can do?
When people log in to their Amazon account, the recommendation function of “People who have bought item X also bought item Y’’ seems simple but effective. Do people realize that it is AI and ML’s work? They are personalizing customers’ experiences base on their purchasing history, browsing history, and other buying habits. The system would like to learn customer’s relevant data as much as possible to provide customers with irresistible products and offers. That’s how Amazon does, and so do other online businesses such as Google, Facebook, Spotify, Youtube, etc. Google knows people’s searching habits, location, language preferences. Facebook knows people’s friends and family groups, and groups of those groups. Spotify pops the customer’s favorite type of songs. Youtube pushes the lasts videos user has been followed. Dr. Know-it-all as an expert in the AI field accepted the interview and explained this phenomenon. In the face of the huge volume of data and a huge number of choices, people will get lost. Through personalization and recommendation systems, people can quickly find what they want. In some cases, it can be said that people have been controlled by AI.
The development of AI
Kai-Fu Lee defined four waves of AI development in his book, internet AI, business AI, perception AI, and autonomous AI. The first wave is about using AI algorithms as the recommendation engine as mentioned above. The second wave is to utilize the data collected in the past in the traditional industries to discover hidden connections that humans tend to ignore and helping humans to make decisions, such as in medical science, financial area, weather forecasting. The third wave is the so-called Online-Merge-Offline (OMO). OMO will integrate the data world and the real world, brings the convenience of the online world into the offline world, and brings the content perceived in the offline world to the online world. An implemented example is the Hema Fresh retail platform under Alibaba Group. Hema Fresh is a new retail format that Alibaba has completely restructured from offline supermarkets. Consumers can buy at the store or place an order on the Hema app. One of the biggest features is fast delivery, within 3km of the store, 30mins home delivery. It’s a fresh food supermarket, a convenience store, and also a restaurant. The supply chain, sales, and logistics links of Hema Fresh are fully digital. And it’s all based on accumulated years of big data and corresponding algorithms from Alibaba. The fourth wave is autonomous AI which combines extremely complex data with machine perception capabilities like Tesla’s autonomous driving technology.
According to Dr.’s analysis, the development of AI is already in the third stage and will enter the fourth wave soon. One example is Tesla’s ‘Full Self-Driving’ beta which was launched in Oct 2020.
AI’s impact on the labor market
Will people’s work be threatened by AI? The answer is yes. Kai-Fu Lee believes that AI and automation will have the technical ability to replace 40-50% of jobs within 15 years. It mainly focusses on the following work and task scenarios,
- Repetitive work, such as washing dishes, an inspection of assembly lines, sewing
- Various interactions with fixed desk books and dialogue content, such as customerservice, telemarketing
- Relatively simple data classification, such as file filing, homework scoring, businesscard screening
- Work in a very narrow field of a certain company, such as an accountant in a certaindepartment
- Work that does not require a lot of face-to-face communication with people such as sorting, data input
What AI cannot do? There are mainly the following aspects:
- Creative work (e.g. medical researcher, AI scientist, writer, artist, entrepreneur). AI is not good at proposing new concepts.
- Complex/strategic work (e.g. CEO, negotiation expert) work that requires knowledge of multiple areas and strategic decisions.
- Sensitive work (e.g. oral surgeon, aircraft mechanic). The robot is still very clumsy when doing handwork.
- Work that needs to adapt to new and unknown environments (e.g. geological surveys). The robot is difficult to adapt to a new environment.
- Empathy/humanized work (e.g. social workers, special teachers). AI does not have human emotional intelligence. People don’t want to “trust” machines and let them handle humanized tasks.
Dr. has the same opinion as Kai-Fu Lee on this question. When asking will the teacher as a career be replaced by AI in the future, he figured out that education is not task-specific but eventually teachers will be less necessary than they are now. It might be the role that will be replaced in the very end. It will be more like in the factory. The factory still has people working with the machine/robots.
AI and business opportunities
As early as 2017, the Finnish government has begun to research how Finland can become one of the world’s top countries in the application of AI technology, and submitted a report on Finland’s Age of Artificial Intelligence to the government. In the report, expertise, high-quality data, and cooperation are seen as key strengths of AI in Finland. Kai-Fu Lee mentioned in his book AI Superpower that successful AI algorithms require three factors: big data, powerful computing capabilities, and excellent AI algorithm engineers. Therefore, Finland does have a lot of business potential in AI technology applications. An example is that Finnish startup company Uniqul is the first one to launch a “face swiping” payment system in the world. According to the AI Business channel from Business Finland, in the past years, Finnish AI technology has been applied in the following areas.
- AI Development & Governance
- Chatbots
- Data & Analytics
- Machine Vision
- Robotic process automation (RPV)
- Neuro-linguistic programming (NLP)
- Marketing & Advertising
- Media & Entertainment
- Forestry
- Retail & E-commerce
- Education
- Talent Management
- Defense, Safety & Aerospace
- Scientific Research
- Automotive, Logistics & Transportation
- Healthcare & Wellbeing
- Financial Services & Legal
- Manufacturing, Process Industry & Mining
- Construction, Building Automation & Real Estate
Although many areas have been covered, as a new startup business, AI technology still has broad development and applications. Either with exploring AI technology and applications or utilize the benefit from AI. In the new AI era, AI can help businesses optimize production processes, sales processes, logistic processes, and customer service. “Data is the new oil”. Who can collect more data in the business model, who will be the future winner? According to McKinsey’s estimates, about 35% of purchases on Amazon come from product recommendations. About 70 percent of users watch is a personalized recommendation on Netflix. In the future, successful AI algorithms will optimize all industries, allowing people to be more creative and have more time to do what they like.
Conclusion
AI is the future. Dr. made some predictions about the development of AI in the next ten to twenty years and its impact on society. One is that people will own fewer cars than now. There will be more electric cars, and when renting a car becomes cheap and fast, people would prefer not to buy one. Second, around the year 2040, human beings will not be allowed to drive especially on highways. Autonomous drive will be the replacement with mature AI techniques. The third is AI will help people live longer and healthy. Re-program genomes are already in use in China for people who have sickness caused by one gene. Fourth, AI will be everywhere even in creative industries. Apparently, the research and application of AI will generate great value. Finding and using the value of AI will bring a broader space for startups. Eventually, it will pay off the investment in knowledge.
Despite this, concerns about AI are also attracting more attention. They mainly involve questions about the singularity, if the government should oversee AI development, AI-related ethics, and data privacy. Dr. gave his opinion about these questions. From different angles, everyone may have their own unique views on these issues. Be ready, the fourth industrial revolution may have come.
References
Interviewee: Dr.Know-it-all
Associate professor at the University of Georgia has a professional background in
astronomy, physics, AI, and English literature. A YouTuber has channel Dr.Know-it-all Knows it all introducing AI, Tesla, and technology-related knowledge. https://www.youtube.com/watch?v=u2nkICOK2J0
Kai-Fu Lee (2018) “AI Superpowers” AI
https://en.wikipedia.org/wiki/Artificial_intelligence
Weak AI
https://www.investopedia.com/terms/w/weak-ai.asp
Deep Learning
https://en.wikipedia.org/wiki/Deep_learning
AI vs Machine Learning vs Deep Learning
https://www.edureka.co/blog/ai-vs-machine-learning-vs-deep-learning/
When and When Not to Use Deep Learning
https://blog.dataiku.com/when-and-when-not-to-use-deep-learning
Hema Fresh
https://baike.baidu.com/item/盒马鲜生
The ROI of recommendation engines for marketing
https://martechtoday.com/roi-recommendation-engines-marketing-205787
Artificial Intelligence from Finland
https://www.magnetcloud1.eu/b/businessfinland/AI_From_Finland_eBook/#p=35
Finland’s Age of Artificial Intelligence
https://julkaisut.valtioneuvosto.fi/bitstream/handle/10024/160391/TEMrap_47_2017_verk kojulkaisu.pdf?sequence=1&isAllowed=y
‘’Data is the new oil”- Clive Humby
https://en.wikipedia.org/wiki/Clive_Humby
What's your reaction?
Post a Comment
