More businesses have deployed AI than ever before, but few are taking steps to ensure they’re using it in a trustworthy and responsible way, according to IBM’s third annual Global AI Adoption Index.
After surveying more than 7,500 organizations in 19 countries, the researchers found 35 per cent have adopted AI, a four-point increase from 2021. Although AI is now more accessible and user-friendly than it was three years ago, the report suggests adoption may be moving more quickly than efforts to define what ‘responsible’ and ‘ethical’ AI looks like.
Here’s a snapshot of how and why organizations are using artificial intelligence today.
Based on IBM’s findings, these are the top five use cases for AI in 2022:
- 1: automation of IT processes
- 2: security and threat detection
- 3: automation of business processes
- 4 & 5 (tie): “marketing and sales” and “business analytics or BI”
The pandemic remains a huge driver of AI adoption, with 31 per cent of organizations citing it as a motivator for deploying AI in 2022. But two other factors are also fuelling AI uptake: its greater availability and ‘useability’:
- 43% cite advancements in AI that make it more accessible
- 37% cite the increasing amount of AI embedded into standard off-the-shelf business apps
- 45% say AI is now better designed to fit the needs of businesses than three years ago
- 41% say AI is now more accessible and easier to deploy than three years ago
It’s clear that vendors have heeded the call to make AI easier to procure and use, and businesses are buying into the technology as a result.
What benefits are businesses getting from their AI tools? Here are their top three answers:
- 54% – cost savings and efficiencies
- 53% – improvements in IT or network performance
- 48% – better customer experience
What really stands out in this study, however, is that organizations are harnessing AI as a tool to cope with the global skills shortage. Paradoxically, AI emerges from this research as both an aggravator of the skills shortage and a partial elixir for it.
When asked to name the biggest barriers to AI adoption, 34 per cent of organizations said “limited AI skills, expertise or knowledge,” making it the top answer. At the same time, AI appears to be playing a crucial role in helping them deal with the overall shortage of skilled labour, even in non-IT parts of their operations:
- 30% say AI helps their employees save time by automating repetitive tasks
- 27% use automation “to address the skills gap”
- 39% say automation is helping them “mitigate labour and skills shortages”
- 23% use AI to “address labour or staffing shortages”
Democratization of AI
AI isn’t just for data scientists anymore.
Although the top two user groups of AI within organizations are still IT professionals (54 per cent) and data engineers (35 per cent), AI has made its way directly into the hands of non-IT users across various lines of business, including:
- customer service professionals (25%)
- marketing professionals (23%)
- product managers (21%)
- sales professionals (21%)
- HR professionals (21%)
- finance professionals (21%)
Cloud and data architecture
The way a company architects its IT for cloud and data is a strong indicator of its AI uptake.
“Companies that have deployed AI are 59 per cent more likely to be using a hybrid cloud or multi-cloud environment than those that have not,” the IBM study states.
By comparison, only eight per cent of organizations who describe their cloud environment as “on-premises” have adopted AI thus far.
There’s also a correlation between sophisticated data architecture and AI adoption. Companies that have deployed AI are nearly three times (283 per cent) more likely to be using a data fabric system than those that have not.
“Companies currently deploying AI are more likely to report their company is using a mix of data and cloud environments that allow them to access their data and run their models whenever they need to,” the IBM study reports. “Companies that have deployed AI are 65 per cent more likely to be using a mix of architectures, including databases, data lakes, data warehouses and data lakehouses compared to those that have not.”
While most organizations have embarked on some sort of path to deploy AI and derive business value from it, they’re still wrestling with how to ensure they’re using the technology ethically and responsibly:
- 74% are not taking steps to reduce unintended bias
- 60% are not developing ethical AI practices
But what constitutes ethical, responsible AI? No one has defined that yet, at least when it comes to creating some sort of global standards. In IBM’s study, more than half of organizations (56 per cent) said the lack of regulatory guidance from governments or industries is a major barrier “when it comes to developing AI that’s explainable and trustworthy.”
AI is now more widespread, accessible, user friendly and business oriented than ever before. As artificial intelligence becomes more woven into businesses, however, the biggest issue to tackle may turn out to be the nuances of humanity rather than the precision of technology.