Chapter 2

Discussion 1
One of the greatest challenges to achieving AGI is measuring general intelligence in machines.
Animals and machines do not have a common evolutionary history, and it is extremely difficult
to compare the general intelligence of two systems. To help quantify machine intelligence,
artificial agents are tested on tasks that mimic animal cognition. These tasks include overcoming
physical obstacles and avoiding negative stimuli, object permanence, planning, and causal
reasoning.
While ELIZA is able to pass the Turing Test by manipulating symbols, this method has a few
limitations. It requires a highly controlled environment, the participants must be concealed from
each other, and the testing must be conducted with reliable means of communication. In addition,
the test may not be appropriate for all types of computing systems, and there may be inherent
limitations that make it unsuitable for the task.
A similar difficulty is faced by knowledge engineers when trying to develop intelligent systems.
They interview experts and try to replicate the experts' knowledge in computer programs, but the
results are mixed (Nishant et al., 2020). One of the first expert systems, MYCIN, was created in
1974. Its ontology consisted of bacteria, diseases, and treatments, but it did not incorporate
patient information.
Alan Turing, an English computer scientist, was one of the first people to become interested in
measuring machine intelligence. Turing reformulated the question and devised a test that tested
for machine intelligence. It was similar to a party game called the Imitation Game. One person
would hide in a room, while another would ask questions to the hiding person. The man would
try to answer as if he were a woman.
Exercise 4
A new five-part video series by McKinsey explores the major drivers of AI and the issues
companies face. In this first segment, we learn about the impact of artificial intelligence (AI) on
Nuance Inc., a company that makes voice recognition software and products.
AI is not a panacea. It has a long history of low ROI, largely because companies focused on
creating AI models without focusing on business problems that truly required its application. The
problem has become more acute as companies move beyond basic AI to more sophisticated use
cases and complex challenges. This requires companies to build more targeted AI initiatives.
While AI may not be a panacea, it can help companies improve the quality of their work. It can
optimize processes, automate tasks, and offer decision support. This can increase productivity
and efficiency, while creating new revenue channels (Chui et al., 2018). AI can also reduce the
workload of human employees, freeing them up to focus on more valuable tasks.
AI has many benefits for businesses, including improved customer experience. For example, AI
can improve marketing by understanding customer needs, matching products to needs, and
persuading people to buy. McKinsey's analysis of AI found that marketing is one of the most
valuable domains where AI can be used.
Exercise 5
A recent McKinsey & Company report on AI drivers revealed several interesting insights. For
example, 36% of executives believe that AI will help improve internal business operations,
engage customers and employees, and obtain insights through data analysis. Meanwhile, 20% of
C-suite executives consider machine learning to be a core component of their business. Some
experts believe that AI could eventually replace most human jobs in the near future. In fact,
McKinsey estimates that artificially intelligent robots will replace 30 percent of the human
workforce globally by 2030.
For example, Merantix, a German company, uses deep learning technology to apply it to medical
problems. It developed an application to automatically detect lymph nodes and other growths in
CT images. Radiologists spend around $100 per hour reading images. If a computer can detect
these changes, it can save the company $250,000 over a period of time.
While AI is a powerful tool, it also poses challenges. Governments must think broadly about the
benefits of AI before implementing new policies. Regulatory crackdowns will only limit
innovation and make AI harder for companies to use. Additionally, AI is subject to bias and
discrimination, which is a serious concern (Balasubramanian et al., 2018). For this reason,
governments should extend existing laws on discrimination to digital platforms. This will protect
consumers and build confidence in AI. AI is becoming a critical part of business. Whether it's for
sales forecasting, marketing, or customer service, companies will need to consider the benefits of
AI.
Exercise 15
Nuance's Dragon voice recognition software is capable of recognizing human speech and
converting it into text. It can also correct mistakes, move through lists, and search text. It also
supports many commands, including the ability to launch applications. However, the program is
not perfect. It requires some training to operate properly, and the user must be careful not to
leave the voice commands turned on when they are not needed.
Dragon voice recognition software from Nuance is one of the most popular options for people
who need dictation assistance. Unlike most other options, it is highly accurate and can dictate
documents three times faster than a person can type. This feature can also be used for quick
sharing of documents through popular cloud-sharing services.

Nuance is one of the pioneers of conversational AI. The company's products include Dragon
Medical One and Dragon Ambient experience, which are designed to help physicians and
clinicians work more efficiently. Additionally, the company also offers Dragon Home, which is a
speech recognition solution for home use (Martínez-Plumed et al., 2021). The company claims
that Dragon Home delivers transcription 3x faster than typing and with optimal accuracy. Dragon
Home can provide a range of features and services to enhance the user's experience. It also
provides a text-to-speech option. This feature is especially useful for people with dyslexia or
sight impairment.
References
Nishant, R., Kennedy, M., & Corbett, J. (2020). Artificial intelligence for sustainability:
Challenges, opportunities, and a research agenda. International Journal of Information
Management, 53, 102104.
Chui, M., Manyika, J., Miremadi, M., Henke, N., Chung, R., Nel, P., & Malhotra, S. (2018).
Notes from the AI frontier: Insights from hundreds of use cases. McKinsey Global Institute, 28.
Balasubramanian, R., Libarikian, A., & McElhaney, D. (2018). Insurance 2030—The impact of
AI on the future of insurance. McKinsey & Company.
Martínez-Plumed, F., Gómez, E., & Hernández-Orallo, J. (2021). Futures of artificial intelligence
through technology readiness levels. Telematics and Informatics, 58, 101525.