Businesses aim to quickly implement AI technology, but encounter numerous obstacles.

• Companies are hindered from moving quickly due to numerous obstacles to the adoption of generative AI.
• Among them: cybersecurity threats, talent shortages, and regulatory delays.
• Keeping up with AI advancements is a continuous task, not a one-time achievement.

Companies seeking to utilize the most recent AI technology are seeking various advantages, such as automating routine tasks, improving data analysis, minimizing human mistakes, and making quicker, more informed decisions.

Barriers to AI adoption hinder companies from deploying the technology as quickly as desired.

A survey of 120 U.S. senior AI/machine learning decision makers, conducted in late 2023 by research and media firm Foundry and technology consulting firm Searce, revealed that less than 40% of organizations have successfully deployed an AI project.

Cybersecurity threats are among the biggest barriers to AI adoption, according to a study by The Foundry/Searce, with 58% of respondents citing data security as a leading concern.

Jake Williams, a faculty member at cybersecurity research firm IANS Research, stated that there is a lack of understanding about the security vulnerabilities of AI applications.

AI apps, particularly those utilizing large language models, introduce a unique set of vulnerabilities that are not well comprehended by most developers and security testers. As a result, some CISOs are cautioning against launching AI projects until there is improved understanding of these risks and better tools for auditing and defense.

To ensure the security of their AI applications, companies should prioritize in-house education and training for data scientists and those with expertise in security and AI, according to Williams. In the future, he predicts that there will be dedicated security training and certifications for AI, but in the meantime, organizations must focus on processes to properly threat model their AI applications and identify unique risks.

AI return on investment

Vrinda Khurjekar, senior director at Searce, stated that many businesses are not considering the unclear use cases for AI and their potential return on investment.

The main reason for poor adoption of AI is the lack of prioritization of a well-qualified use case, according to Khurjekar. If a use case is too impactful, any failures can create doubts across the organization. On the other hand, if a use case has minimal impact, it fails to gain momentum from the rest of the organization.

Striking the right equilibrium between intricacy and influence is vital in the adoption of AI throughout the organization, he emphasized.

To accelerate AI adoption, it is crucial to have a focused approach on which use cases to handle first. This can only be achieved by examining the entire organization, identifying where AI can have the greatest impact, and prioritizing which needs to be tackled first.

Khurjeker suggested that companies should involve all organization members in the council to make it more efficient.

Organizations are eager to incorporate AI into their applications, but lack understanding of its potential benefits, according to Williams. He noted that this rush to adopt AI is driven by a fear of being left behind, without considering its specific use case applicability. This phenomenon resembles the early days of blockchain technology.

The shortage of talent in the AI field is a challenge that many companies are facing, which can hinder their adoption of AI technology.

Rapid advancements in technology are making it challenging for organizations to attract and retain top talent, according to Khurjeker. As a result, teams may struggle to launch AI initiatives or have faulty launches, which can lead to doubts among the broader organization.

Hiring talent and providing training programs to enhance AI proficiency can fortify the talent pipeline, according to Khurjeker.

Another obstacle is the immaturity of low models, which can result in "hallucinations" or instances where AI models produce unfounded information that is not grounded in reality.

Generative AI models are still in their early stages, and their outputs may contain hallucinations. This is particularly concerning for industries where accuracy is crucial, such as healthcare and finance, causing early adopters to exercise caution.

Companies will face a significant challenge in quickly adopting AI tools until the models become more mature, according to Khurjekar.

Regulatory policies and compliance efforts may hinder the adoption of AI.

"Companies in highly regulated industries are hesitant to adopt AI policies due to uncertainty about future regulatory changes, which slows down adoption and results in companies implementing and then unwinding policies if there are significant changes," Khurjekar stated.

Staying current with AI developments is an ongoing mindset shift that requires businesses to look at all processes with an AI-first lens. AI adoption is not a one-time event that companies need to plan for. Khurjekar emphasized the importance of staying up to date with all the latest advancements in AI in order to ensure the success of an AI adoption journey.