When ChatGPT was officially presented to a wide audience in November 2022, the hype around Artificial Intelligence was higher than at any time before. Nevertheless, the development and implementation of AI tools started significantly earlier. And they have already made a long way to their adoption within various industries.


SwissCognitive Guest Blogger: Artem Pochechuev, Head of Data Science at Sigli – “Challenges To The Widespread Use Of AI-Powered Solutions”


SwissCognitive_Logo_RGBCan we say that AI is commonly implemented by corporate and individual users? Well, it depends on the perspective. Without any doubt, the process of AI adoption is gaining momentum but we definitely can’t say that it has already achieved its highest point.

AI adoption: current state and expectations

In order to avoid subjectiveness, let’s take a look at the figures.

According to the latest statistics, today more than 80% of enterprises strongly believe that AI- and ML-powered solutions can help them achieve higher business efficiency, increased revenue, and enhanced customer satisfaction.

In 2022, IBM discovered that 35% of businesses were utilizing AI in their activities while another 42% admitted that they were exploring the potential of such technologies.

The report by CB Insights revealed that in 2023, AI startups managed to raise an impressive amount of $42.5 billion across 2,500 funding rounds. And 48% of this sum was allocated for generative AI.

Thank you for reading this post, don't forget to subscribe to our AI NAVIGATOR!


But what are the most common use cases for artificial intelligence today? Of course, it is quite challenging to get precise data about the popularity of this or that use case in general as reality will greatly depend on the industry and region that we are talking about.

Based on the publicly available data provided by McKinsey. Service operation optimization, the development of new AI-based features, and contact center automation are among the leading use cases.

According to another research, more than 40% of businesses utilize AI and data analytics tools in their digital product development processes or at least for some of their stages.

However, we should admit that the majority of studies aimed at detecting the most highly-demanded AI use cases are conducted among business entities. As a result, a row of social or healthcare organizations turn out to be excluded from the range of respondents.

But as we’ve already highlighted in our previously published articles the role of AI in the work with people with disabilities and their social inclusion can’t be underestimated.

Barriers to AI adoption

Nevertheless, when it comes to AI adoption, a wide range of projects face serious barriers, regardless of their exact type and goals.

Data privacy and security

The core concerns with AI-powered solutions revolve around data security and privacy, given the substantial volumes of data that are necessary for the operation and training of AI systems. The protection against leaks, breaches, and improper usage requires a comprehensive approach. Compliance with data protection regulations such as the CCPA and GDPR demands the implementation of measures, including access controls, encryption protocols, and robust auditing mechanisms within organizational frameworks.

Ethical concerns 

The issues of AI ethics delve into a wide range of complex matters, spanning from breaches of privacy to the propagation of biases and societal repercussions. The core of these problems is hidden in the imperative of holding AI systems accountable. This is required to ensure transparency in their operations and uphold fairness in their decision-making processes. Speaking about the most pressing concerns, we should mention instances of algorithmic bias, driving discrimination against specific demographic groups and highlighting existing inequalities in society.

Lack of diversity in data 

This point is closely related to the previous one. Due to the fact that quite often homogeneous data sets are used for AI training, it can lead to unequal treatment and discrimination. As a result, we can get algorithms that will better suit only specific groups of users, without taking into account the entire population. While in this case, there are a lot of concerns related to race, we’d also like to draw your attention to people with different physical abilities. Today, the majority of AI systems are trained without relying on data that can be relevant to people with various disorders.

Legal issues 

AI uses huge volumes of data. And here is when a large spectrum of legal concerns arise, including intellectual property rights, regulatory compliance, and liability. Moreover, who is responsible for the decisions made by AI? Today there is still no comprehensive regulation for that.

Software malfunction 

The potential risks associated with AI software malfunctions are rather considerable. The range of possible problems varies from inaccurate outputs to system breakdowns and vulnerability to cyber-attacks. Rigorous testing and quality assurance protocols must be implemented at every phase of the software lifecycle to mitigate these risks effectively.

High costs

We should admit that today the development and integration of AI solutions are rather expensive projects. It means that businesses should be ready to cover upfront costs before enjoying the new opportunities that AI can provide. Nevertheless, not all businesses can afford such investments without external support.

But there is also good news. According to the report by IBM, companies have the possibility of reaching a 13% ROI on AI solutions and tools. And it is more than twice higher than an average 5.9% ROI on enterprise-wide initiatives.

Limited accessibility

When we are talking about AI-based products for a wide audience, it is necessary to bear in mind that today not every potential user can get access to the required devices. Let us mention at least one very simple example. Artificial Intelligence can be used to create advanced navigation for people with visual impairments. But to leverage all the available features, it is necessary to buy either AI-powered glasses or special AI-powered devices that can be attached to any pair of glasses.

Tech barriers

The existing infrastructure and systems that are currently used by many organizations are not suitable for implementing AI technology. Its integration with legacy solutions may turn out to be a time- and effort-consuming process.

Lack of AI expertise

The problem is that today we can observe that the number of people with strong AI expertise is lower than the market demands.

And it won’t be an exaggeration to say that the progress in the AI sector is being slowed by a shortage of specialists who have experience in working with NLP, deep learning, and other related technologies. As a result, there is even a risk of wasted investments that could become a consequence of the lack of experts who can deliver AI solutions, provide their support and maintenance, and ensure their smooth implementation.

How is it possible to overcome the existing barriers?

Unfortunately, we do not have a magic wand that can help us immediately improve the situation. Nevertheless, there are some actions that can boost the AI adoption process and make it more seamless for both business and private users. Below, you can find some ideas that look quite efficient in addressing the existing issues. Please, do not consider them as a step-by-step guide or a manual. It is just a list of possible solutions based on our practice and some market research.

First of all, it is highly important to increase social awareness about the potential of AI. If we are talking about businesses it is vital to do it at all levels, including C-level managers and employees who do not participate in the decision-making process. Let’s admit that already now we can see fewer people who voice their fear or aggressiveness regarding the use of AI, especially if we compare today’s situation with what we had at the dawn of the ChatGPT era. The more people know about technologies, the fewer reasons they have to oppose innovations.

Secondly, it is necessary to develop a comprehensive AI adoption strategy across the industries that want to work with such technologies. It is crucial to elaborate common standards and understanding for working with data, its collecting and processing, as well as the introduction of such tools to both internal users and clients. It is necessary to adjust the existing approaches to managing data to new realities. Moreover, for many businesses, it could be sensible to engage experts who already have relevant experience.

And thirdly, we can’t deny the significance of sufficient funding for the development of such projects. That’s why the attraction of private and governmental investors to tech initiatives seems to be an important part of all AI adoption efforts.

So, is the world ready for mass adoption of AI-powered solutins today? It’s difficult to provide a 100%accurate answer. The work is being done in their direction. Nevertheless, a row of quite serious barriers are still here. And, unfortunately, it could be a rather long way to their full elimination.

About the Author:

Artem PochechuevIn his current position, Artem Pochechuev leads a team of talented engineers. Oversees the development and implementation of data-driven solutions for Sigli’s customers. He is passionate about using the latest technologies and techniques in data science to deliver innovative solutions that drive business value. Outside of work, Artem enjoys cooking, ice-skating, playing piano, and spending time with his family.