In recent years, artificial intelligence (AI) has been presented as the next big technological revolution, promising to transform industries and redefine work as we know it.
However, despite the enthusiasm and optimistic projections, the adoption of AI in the business sphere has not reached the expected heights. This is reflected in the valuation of some of the generative AI giants, which has suffered in recent weeks.
Is AI hype? Will the generative AI bubble burst? Let’s explore together the real state of AI adoption in companies and analyze the challenges and opportunities presented in this technological landscape.
A lot of investment, little adoption
Despite significant investments in AI, effective adoption of this technology by companies remains surprisingly low.
According to an article by Sol Rashidi in Forbes, although US companies have invested almost $290 billion in AI, only 5.4% of companies in the United States have implemented this technology in their operations as of February 2024.
This contrast highlights an important gap: while investment in AI is massive, effective implementation lags far behind. In comparison, China, although with lower investments, has achieved a much higher adoption rate, with 58% of its companies using AI.
This suggests that the effectiveness and practical application of AI investments in China are significantly higher than in the United States, where effective adoption is only a fraction of what was expected.
But it all depends on who you ask because according to the study “Unlocking the potential of AI in Europe in the Digital Decade” it is stated that 36% of Spanish companies have adopted Artificial Intelligence, which represents a growth rate of 29% since September 2022. Those that have adopted AI services report multiple advantages, including the streamlining of business processes (79%), cost savings (79%) and greater efficiency (88%). 72% of Spanish companies say that the adoption of AI has led to an increase in their income.
According to Microsoft in its “AI Study: Artificial Intelligence in Spain. How do 277 of the main European companies benefit from AI?” Most Spanish companies have Artificial Intelligence (AI) pilot projects, but only 20% have gone beyond the initial proof-of-concept phases, compared to an average of 32% in the rest of the participating European countries.
If we look at Xataka, which is certainly a reference in the dissemination of many technological topics, including Artificial Intelligence, the use of AI is coming from the bottom up. It is the companies’ own workers who begin to use this technology without the support of their bosses, often paying out of their own pockets for the tools that their company should provide. It stands out that 78% use their own AI tools at work. These data are taken from the Microsoft Work Trend Index study that it presented together with LinkedIn that surveyed 31,000 people from 31 countries (1,000 are Spanish). This study provides other data such as that 64% would not hire workers without knowledge of AI.
Why don’t companies adopt AI?
There are several factors that contribute to this discrepancy between investment and adoption, among which the four main ones have been identified:
- Fear of job replacement: Many workers fear that AI could replace their jobs, which generates resistance to the implementation of these technologies.
- Difficulty calculating business value: The total cost of ownership of AI systems and the difficulty in quantifying their return on investment make companies hesitant to fully commit.
- Technological complexity: Integrating AI into existing technology infrastructures can be complicated and expensive, especially for companies that are already struggling with technical debt.
- Conflicting strategic priorities: Companies must often balance their immediate needs with the exploration of innovative technologies such as AI, which limits the resources available for the adoption of new technologies.
AI is not being used in depth
According to results from Keypoint Intelligence’s “AI Readiness” survey, 94% of companies surveyed report some level of AI usage, with 13% reporting deep enterprise-wide usage.
The first number, although impressive at first glance, hides a more complex reality. Most companies are in the early stages of adoption, using only a few AI features built into existing tools, such as Salesforce, Zendesk, or Microsoft 365.
This limited use suggests that while AI has gained traction, its deep and meaningful integration is still a work in progress, something that can be partly explained by technological inertia.
Technological inertia describes the tendency of organizations to maintain and use old technologies due to various barriers, such as implementation costs, employee resistance to change, and other organizational and cultural factors.
This implies that, despite the availability of new technologies and tools, many companies prefer to continue with their existing systems due to the complexity and risk associated with adopting technological innovations.
Real positive impacts of AI on companies
Despite the challenges, AI is generating positive impacts on companies. The same Keypoint Intelligence report highlights that 61% of respondents have seen increases in efficiency and productivity thanks to AI.
More surprisingly, 53% have reported improvements in the quality of products and services. These benefits underline that AI not only has the potential to speed up processes, but also improve the quality of work performed.
Another interesting finding is that 78% of respondents reported significant improvements in collaboration and knowledge sharing between departments thanks to AI. This often underappreciated benefit suggests that AI can act as a unifying force within organizations, encouraging greater integration.
Is generative AI a bubble?
Generative AI, including models such as GPT-3 and DALL-E, has captured the imagination of the public and media due to its impressive capabilities to create text, images and other content from short descriptions.
Asking whether generative AI is a bubble is relevant in the current context, both for companies and individuals. Well, expectations and enthusiasm often exceed the practical reality of the technology.
As with other AI applications, there is a discrepancy between perceived potential and practical implementation. Many companies are in experimental phases, testing the capabilities of these models without fully integrating them into their daily operations.
This experimentation phase is crucial to avoid costly mistakes and ensure that the technology is used effectively and ethically.
One investment too much? in the future
The sustainability of AI in terms of costs and resources is a relevant concern. Generative AI models require enormous amounts of data and computing power to train, which can be prohibitively expensive for many companies.
Until now, the hardware needed to run large language models is in the hands of a few. Unless more efficient methods are developed for training and deploying these models, their large-scale adoption could be limited.
On the other hand, the real value of generative AI in the business context may not be as immediate as expected. Although impressive from a technical perspective, practical applications that generate a significant return on investment are still in development.
Companies should carefully consider where and how to implement generative AI to ensure that it is not simply a fad, but a useful tool that improves their operations and results.
The future of AI in companies
The true state of enterprise AI adoption is less flashy than the headlines often suggest, but it has significant transformative potential.
AI adoption is happening gradually and thoughtfully, which is positive as companies are taking the time to integrate these technologies effectively and ethically.
The focus must now be on overcoming the identified challenges: addressing fear of job replacement, clarifying the business value of AI, resolving technological complexities, and balancing strategic priorities.
Likewise, it is crucial that companies focus on improving the quality of their data and establishing robust ethical policies to ensure that AI is implemented fairly and effectively.
This post is also available in: