Among the many business turmoil caused by covid-19, here is one that is largely overlooked: artificial intelligence (AI) whiplash.
When the pandemic began to overthrow the world last year, businesses were free to use any tool, including AI, to solve their challenges and serve their customers safely and effectively.To 2021 KPMG Survey Half of US executives, conducted between January 3rd and 16th, said organizations have speeded up the use of AI in response to covid-19. This includes 72% of industrial manufacturers, 57% of technology companies and 53% of retailers.
Most people are happy with the result. Eighty-two percent of those surveyed agreed that AI helped the organization during the pandemic, with the majority saying that AI provided even more value than expected. In a broader sense, almost everyone states that widespread use of AI will make organizations run more efficiently. In fact, 85% want organizations to accelerate the adoption of AI.
Still, emotions are not completely positive. Despite trying to step on gas, 44% of executives believe the industry is moving AI faster than it should be. Even more surprising, 74% claim that the use of AI to help businesses continues to be more hype than it really is. This has skyrocketed in major industries since the September 2019 AI survey. For example, in both the financial services and retail sectors, 75% of executives feel that AI is exaggerated. It has increased from 42% and 64%, respectively.
How do you square these seemingly contradictory views of what KPMG calls AI whiplash? There are some explanations for hype based on our work to help organizations apply AI. One is the simple newness of technology. This can be misleading about what you can and cannot do, how long it takes to achieve enterprise-wide results, and the mistakes your organization can make when experimenting with AI without the right foundation.
While 79% of respondents say AI is at least moderately functional within their organization, only 43% say it is fully functional on a large scale. It’s still common to find people who think of AI as a new machine to buy for immediate results. Also, while AI may have had some success (often a small proof of concept), many organizations have learned that extending AI to the enterprise level can be more difficult. I was surprised. You need access to clean, well-organized data. Robust data storage infrastructure. Subject-matter experts to assist in the creation of labeled training data. Advanced computer science skills. And the approval from the business.
Of course, it’s easy to believe that advocates of AI may have sometimes exaggerated that potential or downplayed the effort required to achieve its full value.
I think basic humanity is at work as to why executives are at odds with the speed of AI adoption. To get started, it’s always easy to believe that the grass on the other side is greener. Also, many may be worried that the industry may be moving too fast, primarily because their organization is not comparable to that speed. It’s easy to succumb to those fears if they experience early-stage problems with AI, especially last year when the world witnessed AI-enabled outcomes such as the record rapid development of the covid-19 vaccine. It may have been.
There is another factor that drives different emotions about the potential of AI. That is, there is no established legal and regulatory framework to guide the use of AI. Many business leaders do not have a clear view of what an organization is doing to manage AI or what the new government regulations will look like. Not surprisingly, they are worried about related risks, such as the development of today’s use cases that regulators could crush tomorrow.
This uncertainty helps explain yet another seemingly contradictory finding from our investigation. Executives are usually skeptical about government regulation, but 87% say government should play a role in regulating AI technology.
Transition from AI whiplash
All organizations need their own playbooks to recover from AI whiplash and optimize their investment in technology, but a comprehensive plan should include five components.
- Strategic investment in data. Data is the source of AI and the connective tissue of digital organizations. Organizations need clean, machine-digestable data labeled to train AI models with the help of subject matter experts. You need a data storage infrastructure that goes beyond functional silos in your business to deliver data quickly and reliably. Once the model is deployed, strategies and approaches for collecting data are needed to continuously adjust and train the model.
- Appropriate talent. Computer scientists with AI expertise are in high demand and difficult to find, but they are essential to understanding AI’s perspectives and teaching strategies. Organizations that cannot build a complete team of scientists in-house need external partners who can fill the gap and help organize the ever-expanding AI vendors and products.
- A long-term business-led AI strategy. Organizations get the most out of AI by thinking about finding solutions to their problems, rather than buying technology and looking for ways to use it. Allows businesses, not IT departments, to drive the agenda. If AI investments tied to business-driven strategies go awry, they are an opportunity to fail quickly and learn. It doesn’t burn fast. But even if the enterprise iterates quickly, the biggest benefits will be realized over the long term and must be iterated in line with a long-term AI strategy.
- Improve culture and employee skills. Few will gain momentum on the AI agenda without the support of the workforce and the culture invested in the success of AI. To win employee commitment, you need to at least provide employees with a basic understanding of technology and data and a deeper understanding of how it can help them and the enterprise. Employee skills are also important, especially if AI takes over or complements existing responsibilities. By embracing data-driven thinking and infiltrating deeper AI literacy into your organization’s DNA, you can scale your organization to success.
- Efforts to use AI ethically and fairly. AI has great expectations, but it can also be harmful if an organization uses it in a way that customers don’t like or discriminates against some segments of the population. All organizations need to develop an AI ethics policy with clear guidelines on how to deploy technology. This policy mandates measurements, checks for data problems and imbalances, measures and quantifies unintended biases in machine learning algorithms, tracks the source of data, and identifies who trains the algorithm. Must be part of. Organizations should continuously monitor the model for bias and drift to ensure that the model decisions are accountable.
Executive goals for AI investment over the next two years will vary by industry. Healthcare executives say their focus is on telemedicine, robotic tasks, and patient care delivery. They say that Life Sciences is considering deploying AI to identify new revenue opportunities, reduce management costs, and analyze patient data. Government officials also said they would focus on process automation and improved analytical capabilities, as well as managing contracts and other obligations.
Expected results also vary by industry. Retail executives predict the biggest impact in the areas of customer intelligence, warehousing, and customer service chatbots. Industrial manufacturers recognize it in product design, development and engineering. Maintenance operation; and production activities. Financial services companies also expect improved fraud detection and prevention, risk management, and process automation.
In the long run, KPMG believes AI will play a key role in reducing fraud, waste and abuse and helping businesses enhance their sales, marketing and customer service operations. Ultimately, we believe AI will help solve fundamental human challenges in a variety of areas, including disease identification and treatment, agriculture and global hunger, and climate change.
It’s a future worth working on. We believe that governments and industries alike have a role to play in achieving that. That is, working together to develop rules that drive the ethical evolution of AI without disrupting the innovation and momentum that is already underway.
Read more on KPMG “Prosperity in the AI world” report..
This content was created by KPMG. It was not written by the editorial staff of MIT Technology Review.
https://www.technologyreview.com/2021/06/21/1026580/navigating-a-surprising-pandemic-side-effect-ai-whiplash/ Navigating Amazing Pandemic Side Effects: AI Whiplash