After a decade of stop-and-go development, artificial intelligence has now begun to provide real, tangible value to the business world. McKinsey published an 80-page report titled “Artificial Intelligence: The next digital frontier?”, which provides a comprehensive analysis of the value that artificial intelligence (AI) creates for businesses.
The report points out that “wide application of artificial intelligence technology will bring great returns to businesses.” This means that the disruptive nature of AI will continue to become more apparent in the future. Governments, enterprises, and developers should all be clear on this point. Moreover, the report raises some interesting points (all of which we will discuss later in this article):
- Outside of the tech industry, AI technology is still in an early experimental phase, and only a small number of companies have deployed AI solutions.
- Early adopters are in a better position to realize the true potential of the technology, while those that come later will struggle to match the pace.
- The dependence of AI on foundations of data and big data means that companies cannot cut any corners, and need to get to work on implementation right away.
- AI brings not only new challenges to companies and developers but also creates new regulatory pressures. For example, workforces need training in response to the developments in machine support, countries/cities need to enter the global labor/capital competition, and also solve moral, legal, and supervisory issues.
Currently, researchers and businesses are focusing on artificial intelligence systems such as robotics and automated transportation, virtual agents, and machine learning (including deep learning and the foundations of several recent advancements in AI technologies). Investments in AI are growing by the day, led by a host of familiar digital giants.
Are Businesses Ready to Leverage AI?
Lately, there has been a lot of talk about the potential and dangers of AI. However, AI is far from a new concept. It has experienced ups and downs throughout history, both expectations and disappointments. Will things be different this time? The answer as per the new analysis is “yes”: artificial intelligence has finally begun to deliver real business benefits. The conditions for a breakthrough are already in place.
Computing capabilities have grown considerably, algorithms have become more refined, and more importantly, a large amount of data has been generated globally, and data, as we know, is the fuel for artificial intelligence. At present, most industry news comes from providers of artificial intelligence technology. Many new use cases are still in the experimental stage. The products on the market are limited, or there are fewer products that businesses can apply instantly and universally. As a result, analysts’ have two distinct opinions: some are optimistic about the potential of artificial intelligence, and some are still very cautious about their economic benefits. This inconsistent concept leads to huge differences in the size of the market.
Driven by technology giants, investment in AI is increasing, but commercial applications are still falling behind.
Technology giants are investing billions of dollars in AI. They see the future direction of AI technology — robust computer hardware, increasingly complex algorithmic models, and vast amounts of data — all of which have already been realized in part. In fact, the internal investment of large companies occupies a major position in the field of artificial intelligence.
AI Increases Profits and Promote Industry Transformation
Though AI has developed rapidly in the recent years, the subsequent adoption is still in its infancy. This makes it challenging to assess the potential impact of artificial intelligence on companies and industries. What happened to the companies that have already invested in AI? McKinsey has found early evidence of large-scale adoption of AI bringing lucrative returns.
It reviewed a large number of case studies in five industries to study how AI can transform some business activities and bring potential fundamental changes to other businesses. These cases demonstrate how AI is shaping different functions across the entire value chain and different industries. These cases also have a wide range of impact on stakeholders, such as multinational corporations, start-ups, governments, and community organizations.
Industry Case Studies Demonstrate the Disruptive Potential of AI
The study includes five case studies to provide an understanding of the wide-scale application of AI in the commercial field. These case studies show how AI influences specific behaviors through multiple forms. The study covers retail, electricity, manufacturing, healthcare, and education industries. Types of companies include private, public, and social enterprises, including heavy asset operations from labor-intensive industries to B2B.
If it is to meet expectations, artificial intelligence needs to play an actual role in the field of economics to significantly reduce costs, increase profits, and increase asset utilization. The study includes a classification of the ways in which AI can create value in four areas:
- Enable companies to better plan and forecast demand, optimize research and development, and increase resources
- Improve ability to produce goods and provide services at a lower cost and higher quality
- Deliver products to customers at the appropriate price and with the correct information
- Provide personalized and convenient user experience
The creation of value in these four fields depends on specific use cases. Many organizations have already discovered or deployed solutions around these use cases (we have listed these AI use cases at the end of this article). Similarly, these use cases have different relevance from industry to industry, meaning that there are numerous opportunities to use artificial intelligence in the planning and production layers.
Additionally, when machine learning can bring valuable benefits to all industries, some particular technologies may have unique commercial application in specific sectors. For example, robotics in sales and manufacturing, computer vision in the medical industry, and natural language processing in the education industry.
Recognizing the True Potential of AI
It is time for enterprises, developers, and governments to recognize the true potential of AI. Although artificial intelligence has the potential to reshape society as a whole fundamentally, we are still unsure how the technology will develop. For enterprises, governments and workers, this uncertainty means that we have to “wait and see.” However, we still consider it necessary to take positive and clear actions to make the most of the emerging opportunities while tackling risks.
For many businesses, this means that they need to accelerate the digitization process and ensure that they deploy AI tools efficiently. Because AI integrates large amounts of high-quality data into automated workflows, the influence of that data is also growing. AI is not a shortcut to the foundation of digitization; rather it is a powerful extension of that digital foundation.
Developers play a crucial role in helping companies realize the potential of technology. The problem that developers need to consider is that AI products need to solve practical business problems. Instead of developing interesting solutions, they have to solve real-world problems on a large scale.
The governments and workers need to prepare for the transformational changes that AI can bring in the future. There is a need to rethink the public education system and employee training problems to ensure that the skills that employees have are complementary to machines, rather than competing with them. Furthermore, regions or countries wishing to establish a local AI ecosystem must join the global competition for AI talents and investment.
The unresolved legal and ethical issues for the society as a whole are likely the most significant obstacles to realizing the true benefits of AI.
What Factors Are Necessary for a Successful AI Transition?
Use cases/Sources of value
- Browsing use cases
- Express commercial needs clearly and create commercial use cases
- The data ecosystem
- Breaking data silos
- Determine the integration and pre-analysis level
- Identify high-value datac
Technology and tools
- Identify the right artificial intelligence tools for the job
- Fill gaps in capacity using partnerships, mergers, or acquisitions
- Adopt a flexible “trial and error” approach
- Integrate artificial intelligence into existing workflow
- Optimize human-machine interfaces
- Open up organization culture
- Adopt a culture of openness and collaboration
- Trust in artificial intelligence
- The workforce can relearn skills according to demands
Regarding career distribution, there are only a few occupations that automation will completely replace. The analysts at McKinsey claim that in 60% of the occupations, automation will be possible for only 30% of the work. From a geographical perspective, America and China are leading the world in AI, and Europe is, unfortunately, lagging behind.
Challenges for AI Adoption
The rise of artificial intelligence poses a wide range of issues to government and society. In this report, McKinsey not only pointed out some of these issues but also proposed some ways to solve them. Our progress on solving these issues is critical to realizing the potential benefits and avoiding the risks of artificial intelligence.
Encouraging Wider Use of Artificial Intelligence
Current AI applications are concentrated to industries that are already at the forefront of new technologies. Extending the application scope of AI to support new technological fields, especially for smaller companies, is essential to ensure the growth of productivity and economic development, and can ensure a healthy and competitive market. The application of AI across industries can also help in balancing the wage levels of different industries. AI can increase productivity and thereby increase wages. Instead of restricting AI to frontier companies and employees who are already at the top of the income pyramid, a wider scope of application will push the benefits of AI to more companies and their employees.
Solving the Problem of Employment and Income Distribution
The AI-driven automation revolution will deeply affect the work and wages of people everywhere. In the McKinsey survey, an overwhelming majority of companies stated that they didn’t believe that AI poses any threat to employment. However, it is clear that some skills in some occupations will fail to meet the requirements of the future. The government may have to rethink the model by which they provide social services. Of course, there will be several different approaches to solving these issues, including adjustments to the sharing of labor, negative income tax, and global basic income levels.
Addressing Ethical, Legal, and Regulatory Issues
AI raises a series of ethical, legal and regulatory issues. For instance, there is a common real-world risk of prejudice seeping into training datasets. Due to racial, gender, and other prejudices, the real-world data that machine learning algorithms thrive on is also unavoidably full of discriminatory features. Eventually, this can affect AI systems as they are at risk of developing these biases in the training process.
These issues become even more intense as the prejudices become more internalized. At the same time, the public may be skeptical of these algorithms themselves. Since the ethical opinions of the programmers may be coded into the algorithm, which parts of the algorithm do people have a right to know about? Who is responsible for what the AI outputs? This has led to calls for algorithmic transparency and accountability.
Another issue is that of privacy. Who owns the data? What measures do we need to take to protect highly sensitive data (like medical data) without harming data availability? Organizations and institutions that are working to address these issues include Partnership on AI, OpenAI, Foundation for Responsible Robotics, and the Artificial Intelligence Ethics and Governance Foundation.
Ensuring Availability of the Training Data
A large amount of data is crucial to training an artificial intelligence system. Opening up public-sector data can stimulate private-sector innovation, and setting up common data standards can also be quite helpful. In the United States, the Securities and Exchange Commission forced all listed companies to disclose their financial statements in XBRL (Extensible Business Reporting Language) format in 2009 to ensure that the public data is machine readable.
Deploying Artificial Intelligence in Government
Artificial intelligence has great potential for the public sector. Its ability to enhance planning, goal setting, and personalization of services, makes it essential to increasing the quality and efficiency of government services. In the appendix to the report, the report explores the future of AI technology in two major public areas — medical and education.
AI Use Cases
The report details five specific use cases in the appendix section. McKinsey gave a visual description of three of these use cases. We have summarized these AI use cases in the section below:
Retail Marketing and Supply Chain
- Face recognition software, machine learning, and natural language processing can enable virtual agents to provide a variety of services.
- Machine learning for personalizing consumer recommendations and matching.
- Deep learning supports the ability of computer vision to identify consumer packages. Adding in sensor data, artificial intelligence makes automatic billing and payment possible.
- Drones powered by deep learning can make deliveries while automatically avoiding obstacles and handling situations when the recipient is not present at the delivery address.
- Interactive screens and desktops can use computer vision and deep learning to identify products, recommend related products, and supplement consumer profiles.
- An automatic shopping cart may be capable of following shoppers in the store, and then delivering them directly the shopper’s car, or pass them to a robot or drone for automated home delivery.
- With machine learning, stores can update and optimize prices in real time based on competitors’ prices, weather, and storage to maximum profits.
- AI enhanced robots can continuously track warehousing, identify empty shelves, and replenish them. Other robots can assist in packaging items in the warehouse.
Power Generation and Distribution
- AI can make the grid more intelligent and trim the number of necessary power stations.
- The data collected from sensors could allow the machine learning system to adjust the output of the power plant in real time.
- Machine learning can predict peak power demands and maximize the efficiency of intermittent renewable energy use.
- An intelligent network cable can then adjust the electric current in real time to improve grid load.
- Drones and miniature robots could detect and predict equipment damage without shutting down the entire line.
- Power companies can record data automatically to reduce the necessary number of technicians.
- The ability to receive data in real time can save effort and time in manual inspections.
- Virtual assistants can help users handle transactions and provide early warning of wrong orders.
- Smart electric meters can automatically adjust power consumption data based on factors such as usage and weather.
Medicine and Healthcare
- AI can help in providing faster diagnosis, better medical planning, and more comprehensive medical insurance.
- Machine learning applications can analyze health data collected on wearable devices, provide fitness advice, and predict a user’s risk of disease.
- Automatic sensors can help patients to monitor physical signs without doctors and nurses.
- Diagnostic tools can diagnose diseases more quickly and accurately with the help of medical data and records.
- By using medical and environmental factors to predict links between patient behavior and likelihood of illness, AI can optimize hospital operations, staffing, and inventory management.
- AI tools can analyze the patient’s medical records and environmental factors to identify patients who are at risk and provide them with preventative health plans.
- Virtual assistants can help patients quickly find the right doctors, save waiting time, and enhance the medical experience.
- Personalized treatment plans, assisted by machine learning tools, can allow patients to recover more quickly. AI-driven Big Data health analysis can reduce hospitalization time.
- AI insights from public health analysis can help taxpayers reduce hospital treatment costs by encouraging providers to manage patient health.
We have discussed how AI is solving real-world business problems and is gaining traction as a ubiquitous technology for businesses of all sizes and scale. The need of the hour is to recognize AI’s unique value proposition, make it an integral part of business strategy, and realize its transformational benefits.
Read similar articles and learn more about Alibaba Cloud’s products and solutions at www.alibabacloud.com/blog.