By Ken Jaroenchisakon Dana Burkart
AI is rightfully getting a lot of attention these days. Unfortunately, some of that attention consists of hyperbole and myths. It’s true that AI is going to transform business and daily life, too — it’s doing so already. But it’s important to understand how that transformation is taking place and how it’s likely to take place in the future.
To encourage open-mindedness and some appropriate excitement without succumbing to fear or complacency, here’s a quick look at five popular myths about AI.
Myth #1: Wherever AI is deployed, it will take people’s jobs.
In reality, the impact of AI will vary from industry to industry and from use case to use case. We predict that AI will make people more productive, complementing employee’s existing skills and enabling them to take on bigger challenges with a high degree of success.
There’s already evidence for AI making employees happier and more productive.
For example, in a story reported by NPR’s Planet Money, AI helped a company’s help desk improve the productivity and effectiveness of its agents by suggesting responses to customer queries. Customer satisfaction increased and employee turnover, which averages about 60% in the call center industry overall, declined. Employees felt less burned out, and they stayed in their jobs, helping customers more quickly and effectively than had been possible before.
A recent study by Goldman Sachs found that about two-thirds of U.S. occupations could be automated by AI in some way. About a quarter of those professions might be replaced — mostly clerical and secretarial roles, according to the World Economic Forum (WEF) — but AI is expected to create new jobs as well. The Goldman Sachs authors cite research by economist David Autor, who showed that 60% of employees today work in occupations that didn’t exist in 1940. In other words, 85% of job growth since 1940 involves new occupations created by advances in technology. It makes sense to extrapolate that AI will lead to new occupations that are difficult to imagine today.
AI will also complement, rather than replace, many jobs. In fact, complementing jobs is likely to be the predominant result of AI deployments. The Goldman Sachs study notes, “Although the impact of AI on the labor market is likely to be significant, most jobs and industries are only partially exposed to automation and are thus more likely to be complemented rather than substituted by AI.”
Myth #2: We can’t take advantage of AI without jeopardizing the privacy of our data.
Data privacy is top of mind for all organizations. What happens if your employee types sensitive company data into an AI chatbot? That data becomes the property of the parent company that owns that chatbot, compromising your organization’s intellectual property.
But there are other AI solutions that do protect the confidentiality and security of company data. There are products, for example, that don’t share the contents of their Large Language Model (LLM) beyond the community of authorized users in a company. Instead, these products keep confidential data in a “data clean room,” while allowing companies to analyze that data in conjunction with larger data sets of non-private data.
Intelligent recommendation capabilities like Boomi Suggest help organizations decide what data to connect when integrating systems and automating workflows. Boomi Suggest draws only on anonymized metadata, learning from integration processes that are built by thousands of companies using Boomi, not the data that those integrations process.
Our advice? Investigate any AI solution’s data security privacy policies before entering proprietary data into it. And feel confident using AI solutions that do protect your company’s data and provide flexibility to opt out of the technology, so you can take advantage of the tremendous power of AI without jeopardizing security or compliance.
Myth #3: We’re still building our data core, so we’re not ready for AI yet.
In a recent article in Fortune, Julie Sweet, CEO of Accenture, points out that some companies don’t feel that they’re ready to take on innovative AI projects, because they’re still building their data core. She cites the example of a healthcare provider that probably shouldn’t be experimenting with a generative AI chatbot for patients if they don’t first have a single view of the patient.
While it is true that you cannot build your own AI models without a data core, it does not mean that you cannot take advantage of AI solutions to innovate and make strides toward your organizational goals. There is no one-size-fits-all approach. AI can be leveraged today in different ways by different groups of users.
For example, AI can help improve master data management for data stewards through discovering, cataloging, and managing their data. And what if we could lighten IT workload by giving conversational AI tools to citizen integrations, employees who have more business than technical expertise, so they could easily get started with a first draft of their integration process by using everyday language? In fact, that’s one of the premises of our Boomi AI offering.
The bottom line is that organizations don’t have to wait to become AI-ready and provide employees access to much-needed productivity tools. Given the pace of AI innovations and new product introductions these days, a company that declares a moratorium on all AI projects risks finding itself lagging behind competitors.
Myth #4: AI is promising, but it’s too early to apply.
Is it too early if a majority of enterprises are already adopting AI? McKinsey reports that in 2018, 40% of companies were investing at least 5% of their IT budgets in AI. By 2022, that number had risen to 52%. And about two-thirds of companies surveyed plan to increase their AI investments over the next three years.
AI is already delivering impressive results in many areas of business today. AI is:
- Analyzing business data and making predictions about sales and inventory
- Analyzing shipping data and offering real-time supply chain guidance
- Analyzing website visitor data and recommending products for specific customers
- Analyzing customer interactions and performing sentiment analysis
- Analyzing chatbot messages from customers and suggesting responses
- Analyzing network traffic to detect and characterize security threats
- Helping software developers generate code five to ten times more quickly
In fact, AI itself is so versatile, it’s difficult to think of an area in enterprise IT and business functions where AI doesn’t have some useful application already. And those applications of AI are paying off for companies across multiple departments and areas of operation.
That same McKinsey survey found large percentages of companies are reporting positive revenue results from AI adoption. The biggest revenue effects are found in product and/or service development and marketing and sales functions, according to 70% of respondents. On the cost side, to no surprise, 52% of respondents have seen the biggest benefits in supply chain management because there are a high number of repetitive tasks that could be automated.
Across business functions, on average, 63% said they have seen revenue values while 32% have reported cost benefits from AI adoption. Clearly, it’s not too early to apply AI. The question, really, is where to apply it and how.
Myth #5: AI is only for developers with deep technical knowledge.
Many of us, technical or not, are already using AI every day. Thanks to natural language interfaces, we interact with AI whenever we use a digital voice assistant like Siri or traffic apps that take into account road conditions, traffic jams, and weather.
The commercialization of products such as OpenAI’s ChatGPT and DALL-E makes it easy for even non-technical employees to experiment with AI.
ChatGPT was initially described as a chatbot, but now is being used to generate written content, accepting a simple prompt such as, “Write a business plan for a managed service provider offering IT security services to financial services companies,” to automatically generate a (typically) well-written response drawing on vast amounts of examples collected from the internet. A more visually oriented application of AI, graphics app DALL-E creates artwork based on instructions such as, “Paint a picture in the style of John Singer Sargent of David Ortiz standing at bat in Fenway Park.”
The ease of use — and rising popularity — of these tools demonstrates that generative AI can be made available to just about anyone.
Ideally, that convenience and the ease of use of these tools should never undermine IT policies and oversight. As companies develop AI strategies, they should look for ways to harness the full power of AI while ensuring that AI never creates new security or compliance risks.
We recommend that companies look for enterprise-grade AI solutions built to give organizations the visibility and control they need to ensure that AI is never misused. That way, companies can realize the full benefit of these new tools without increasing risks.
Learn more about how Boomi helps organizations become AI-first enterprises through intelligently building integrations and automating workflows at boomi.ai.