Artificial intelligence: improving man with machine
A subfield of computer science, artificial intelligence (AI) develops computers that can do things traditionally done by people. Its solutions can sense or perceive the world and collect data; understand the information collection; and act independently – all underpinned by the ability to learn and adapt over time.
Augmented reality, virtual reality, natural language question answering, machine learning, autonomous vehicles – artificial intelligence powers most of the innovations that dominate today’s conversations about the industries of tomorrow.
As those innovations become capable of making significant productivity impacts, business is starting to take note. Accenture has found that the number of AI start-ups in the United States has increased 20 times over the past four years.¹ Jeff Dean, a senior fellow at Google, said in 2015 that he believed the pace of AI’s advancement was “actually speeding up”.²
Narrative Science has reported that 62% of enterprises will be using AI by 2018.³ In the longer term, Accenture’s own research suggests AI could double annual economic growth rates in 2035, primarily by increasing labour productivity by 40%.⁴
AI’s potential to boost economic growth
Source: Accenture and Frontier Economics
As of 2015, the transport and automotive sector held the greatest share of an AI market that Markets and Markets predicts will grow to $5.05 billion by 2020.⁵ In that period, AI is set to grow fastest in healthcare, where it can make treatments more efficient, and improve diagnosis, patient care and drug discovery. Retail, professional services, and oil and gas are among other important sectors.
Applications to industry
Automotive and Logistics
Around the world, road traffic accidents kill 1.25 million people a year, injure 20 million to 50 million, cost $518 billion and are predicted to be the fifth leading cause of death by 2030. Crash-avoidance technology in assisted-driving vehicles is already lowering accident rates. When AI-driven autonomous cars reach critical mass, those rates are likely to plummet.
The insurance aspects of this transformation are not fully clear. What is clear, however, is that a liability shift will take place. Instead of driver behaviour, coverage will be placed on a car’s manufacturer, the software designer, device maker, map producer, the company that made the sensors in the highway or the vehicle, the operator, the passenger or the vehicle’s owner.
Case study: Otto – a self-driving truck company
Recently bought by Uber for $680 million, Otto is testing self-driving trucks on roads throughout Northern California.⁶ For now, the robot truckers are only taking control on highways, but the goal is a more automatized, totally functional rollout by 2020.⁷ Trucking is a $700 billion industry in the United States, and a third of costs go to compensating drivers.⁸ Given the size of the industry, it is likely self-driving trucks will be on the market before self-driving cars.⁹
Already using it behind the scenes for supply-chain cost optimization, retailers are set to deploy AI on the customer-facing front line, where its deeper understanding of consumer behaviour can increase revenues.
In physical retail stores, AI-enabled digital assistants will seamlessly and automatically find, order and deliver the ideal option to customers, and thus help satisfy the growing consumer expectation of instant gratification.
Case study: Pepper – the humanoid robot
In Japan, Nescafe deployed 1,000 Pepper robots to retail appliance stores to help consumers pick out Nespresso coffee machines. Japanese telco SoftBank also experimented with Pepper earlier this year, opening a pop-up mobile phone store run entirely by robot sales clerks that speak 19 languages, recognize facial expressions and learn from conversations. It plans to continue testing the robots and deploy them across Tokyo to serve foreign visitors to the 2020 Olympic Games.
The volume of data produced by healthcare organizations has increased tremendously. Driving this increase has been developments such as the digitization of clinical information through the implementation of electronic medical records (EMRs), the generation of significant amounts of real-time data by billions of connected devices,¹⁰ and lower-cost access to genomic information – not to mention the wealth of information captured on the internet.
This information is feeding next-generation analytics technologies such as big data, cognitive computing and machine learning to, for example, improve the delivery of cancer treatments, personalize medical interventions, predict chronic diseases and drive behavioural change.
Case study: IBM – using AI and machine learning to help physicians¹¹
Medical images make up at least 90% of all medical data, according to IBM researchers. Aiming to help clinicians extract insights from imaging data, the computing giant’s Watson Health and Merge Healthcare arms recently partnered with the Radiological Society to demonstrate some AI-driven solutions. Watson Health has developed cognitive tools for peer review, data summarization and physician support, as well as the MedyMatch Brain Bleed application, designed to help emergency-room physicians diagnose stroke or brain bleed by identifying relevant evidence in patient records. Merge’s Marktation augments the work of physicians by raising image-reading speeds and accuracy. It also has a cloud application for eliminating common causes of errors in medical imaging, and a Lesion Segmentation and Tracking Module.
There is an unprecedented opportunity to harness the power of artificial intelligence to augment humans’ abilities to ‘do’, ‘think’, ‘learn’ and ‘feel’. By automating routine tasks, AI frees humans to focus on solving higher-order problems.
Case study: BakerHostetler – ROSS, the artificially intelligent lawyer
Law firm BakerHostetler recently employed its first AI lawyer, ROSS, for its bankruptcy practice, where 50 human lawyers currently work. Built on IBM’s cognitive computing platform Watson, ROSS is designed to read and understand language, postulate hypotheses when questioned, conduct research, and generate responses (along with references and citations) to back up its conclusions. ROSS also learns from experience, gaining speed and knowledge the more lawyers it interacts with. It also constantly monitors current litigation so that it can notify colleagues about recent court decisions that may affect their cases.¹⁴
Unlocking value to society
Accenture’s work with the World Economic Forum on the Digital Transformation Initiative has uncovered a variety of possible societal impacts stemming from the adoption of AI by different industries.
Integrating predictive intelligence technologies into the design of smart cities promises to improve public safety.¹⁵ Singapore has already deployed sensors and cameras to monitor public spaces, but its Smart Nation programme has also raised concerns about Big Brother-style mass surveillance.
In healthcare, AI is helping reduce the time it takes to bring new drugs to market.
The impact of AI on bringing new drugs to market
Source: ‘Turning artificial intelligence into business value. Today.’ Accenture¹⁶
There are wider, cross-industry questions to be answered too:¹⁷
- Unemployment. As machines take over more mundane tasks, how do we prepare displaced humans to fill the roles created by new technologies?
- Inequality. How should the wealth created by machines be distributed?
- Humanity. How do interactions with machines affect our behaviour?
- Artificial stupidity. How can we guard against mistakes?
- Racist robots. How do we eliminate AI bias?
- Security. How do we keep AI safe from adversaries?
- Evil genies. How do we protect against unintended consequences?
- Singularity. How do we maintain control over complex, intelligent systems?
- Robot rights. How do we treat AI humanely
AI has the potential to drive significant productivity increases that improve the quality of lives around the world, but its development and use is not without risk. As such, it must be implemented responsibly.
Case study: establishing a cross-industry partnership to address AI-related challenges¹⁸
In 2016, Facebook, Amazon, Alphabet, IBM and Microsoft launched a partnership to research and promote best practices around AI. This partnership can be viewed as an example of self-regulation and is intended to provide a formal structure of communication. All the parties have made a strong commitment to keep the public informed about the latest developments in AI research. They also plan to expand the partnership by including non-profit organizations, ethicists and activists, and, in this way, accommodate diverse sources of expertise.
From our DTI research, we have identified several technologies (3D printing, artificial intelligence, autonomous vehicles, big data analytics and the cloud, the Internet of Things and connected devices, and robots and drones) that are having major impacts across the 13 industries analysed to date. This article is one of a series looking at how each of these technologies is transforming business and wider society.
1 Accenture, Technology Vision: People First: The Primacy of People in a Digital Age, 2016.
2 Clark, Jack, “Why 2015 was a breakthrough year in artificial intelligence”, Bloomberg Technology, 8 December 2015
3 Press, Gil, “Artificial intelligence rapidly adopted by enterprises, survey says”, Forbes, 20 July 2016.
4 Purdy, Mark and Paul Daugherty, Why artificial intelligence is the future of growth, Accenture, 2016
5 MarketsandMarkets, Artificial intelligence market worth 5.05 billion USD by 2020 [Press release], PR Newswire, 4 February 2016
6 Prigg, Mark, “Self-driving trucks set to take American roads by the end of the year”, Daily Mail, 5 August 2016
7 West, Darrell M., Moving forward: Self-driving vehicles in China, Europe, Japan, Korea and the United States, Center for Technology Innovation at Brookings, September 2016
8 Kitroeff, Natalie, “Robots could replace 1.7 million American truckers in the next decade”, LA Times, 24 September 2016
9 Muoio, Danielle, “A driverless future is coming – but it won’t start with self-driving cars”, Business Insider UK, 26 September 2016
10 Evans, Dave, The Internet of Things: How the next evolution of the internet is changing everything, Cisco, 2011
11 IBM, IBM unveils Watson-powered imaging solutions for healthcare providers [Press release], 29 November 2016
12 Sanghani, Radhika, “Technology leads to 84pc increase in office productivity”, The Telegraph, 22 October 2013
14 Addady, Michal, “Meet Ross, the World’s First Robot Lawyer”, Fortune, 12 May 2016
15 World Economic Forum and Accenture, Digital Transformation of Industries: Aviation, travel and tourism, 2017.
16 Bataller, Cyrille and Jeanne Harris, Turning artificial intelligence into business value. Today, Accenture, 2016
17 Bossmann, Julia, “Top 9 ethical issues in artificial intelligence”, World Economic Forum, 21 October 2016