Visit cnbccouncils.com/tec to meet the CNBC Technology Executive Council.
Numerous challenges hinder the swift adoption of generative AI by businesses. These obstacles range from security threats and skill shortages to regulatory delays. Implementing artificial intelligence is an ongoing process that necessitates staying abreast of the latest advancements.
Businesses seeking to leverage cutting-edge AI tools anticipate various benefits, including task automation, enhanced data analytics, reduced human errors, and improved decision-making speed and quality.
However, the road to successful AI implementation is fraught with obstacles, impeding businesses from embracing the technology at their desired pace. A study conducted in late 2023 by Foundry and Searce surveyed 120 senior AI/machine learning decision-makers in the United States and revealed that less than 40% of organizations had effectively executed an AI project.
One significant barrier to adoption is the escalating security risks associated with AI deployment. Data protection concerns were cited as a major impediment by 58% of respondents in the Foundry/Searce research. Jake Williams from IANS Research highlighted the lack of awareness regarding security vulnerabilities in AI applications, especially those utilizing large language models. He emphasized the urgent need for a deeper understanding of these issues and enhanced auditing and security tools before expanding AI initiatives.
Williams suggested that businesses focus on enhancing their understanding of AI functionality. He predicted a rise in specialized training and certifications for AI professionals in the coming years. Given the current immaturity of tools in this domain, companies must prioritize developing robust risk modeling techniques for AI applications.
Return on Investment for AI
Unclear AI use cases pose another challenge to adoption. Vrinda Khurjekar, a senior producer at Searce, noted that many businesses struggle to identify high-impact AI use cases that offer substantial returns on investment. Selecting overly complex use cases can lead to failures that sow doubt across the organization, while opting for low-impact scenarios may fail to garner sufficient support.
Khurjekar stressed the importance of striking a balance between impact and complexity when selecting AI use cases. Establishing AI governance frameworks can help prioritize use cases effectively, accelerating the adoption of AI within organizations.
To address the talent shortage in the AI field, businesses must invest in proactive talent acquisition strategies and provide training programs to upskill existing employees. Without a skilled workforce, AI initiatives may falter or result in subpar outcomes, undermining organizational confidence.
Another obstacle is the nascent stage of AI model maturity, particularly generative AI models, which may produce misleading or inaccurate results known as “hallucinations.” Khurjekar cautioned that industries reliant on precise data, such as healthcare and financial services, must proceed cautiously until AI models mature further.
Furthermore, compliance efforts and evolving governmental policies pose challenges to AI implementation. Regulatory bodies are still formulating guidelines for widespread AI usage, causing uncertainty among businesses, especially in highly regulated sectors. Businesses await clearer regulatory frameworks to avoid investing in AI solutions that may later require costly revisions.
Staying informed about AI advancements is crucial for businesses navigating the complexities of AI adoption. Khurjekar emphasized the need for a continuous learning mindset, viewing all processes through an “AI-first” lens to ensure successful implementation and alignment with the latest industry developments.
