Risks and challenges of artificial intelligence for business
Businesses keen to put artificial intelligence (AI) technologies to work to improve their profitability, productivity, and business results, need to be mindful of some of the pitfalls. As with any emerging tech, and notwithstanding the tangible and monetary business benefits that artificial intelligence brings, artificial intelligence has various risks and challenges which could hinder business adoption.
While these risks shouldn’t be ignored, it is worth keeping in mind that advances in AI enable businesses to personalise customer experiences, tailor products and services, and exploit growth opportunities with speed and precision that’s never been possible before.
72% of executives consider AI to be a significant business advantage
PWC
On this page:
Challenges of artificial intelligence
The adoption of artifical intelligence has profound implications for many businesses. As with any developing technology being adopted for business, artificial intelligence (AI) has its associated risks, challenges, barriers and disadvantages:
- Data availability – Within most organisations, data is presented in isolation, within silos, across business divisions or is inconsistent and of poor quality. Such data presents significant challenges for companies wishing to extract value from AI. To overcome this, businesses should seek to identify a clear strategy from the outset to ensure that AI pre-requisite data can be derived from data sources in an organised and consistent way.
- Lack of talent – Organisations wanting to leverage the benefits of AI, may face a shortage of employees, both in-house and potential, who possess the necessary skills and experience to utilise the potential of artificial intelligence fully. A lack of AI skills, along with limited availability of technical staff with the required expertise, means that businesses need to consider investing in employee training to implement and operate AI solutions effectively.
- Cost and implementation time – The cost of implementation of AI solutions, both from a time and economic perspective, is a critical factor in choosing to execute this type of project. Businesses which lack internal skills or are not familiar with AI technology. As such, they must rely on outsourcing of both implementation and maintenance to overcome integration challenges and ensure usability and interoperability with other systems and platforms.
Risks of artificial intelligence in business
The emergence of AI is paving the way for a whole new set of operating and business models, with many businesses already leveraging artificial intelligence technologies in some form. However, while AI may be intelligent, it’s still a machine. As AI continues to bring incremental benefits to businesses, it’s not without its inherent risks:
Safety & social manipulation
All technology, no matter how intelligent, can malfunction. Instances of where artificial intelligence has failed, include when Facebook’s AI chatbots started interacting with one another in a new language no one else could understand, or the hijacking of Microsoft’s Twitter chatbot.
As AI solutions diffuse throughout business operations, affecting everything from customer service and sales to back-office automation, examples such as those above give rise to concerns that artificial intelligence could do something harmful which may be detrimental for the business deploying it.
Trust & security
Stemming from complex algorithms and vast data, for some, AI can be a challenging concept to understand. For those who do not have the knowledge of these algorithms, artificial intelligence may be a foreign concept, due to which, they lack trust in the technology.
However, such a lack of trust may be warranted. Without assurance over the technology, how can businesses ensure that the AI algorithms are making decisions that align with corporate or moral values? Or how do businesses retain control over technology which is design to learn itself? Without formal regulations or guidelines in place, many businesses still lack trust in AI.
Job losses & organisation restructuring
According to the World Economic Forum, artificial intelligence technologies will replace more than 75 million jobs by 2022. Job loss concerns related to artificial intelligence has been a subject of numerous analysis, research and academic studies.
As humans and machines collaborate more closely, companies will aim to improve their margins by increasingly automating their workflows, particularly the lower-level, repetitive tasks, potentially leading to job losses. However, research indicates that AI doesn’t always perform best on its own. While job losses may be inevitable, organisations will need to replenish their workforce with new skills and expertise.
Ethical concerns of AI
In the absence of an international set of standards to identify the boundaries of artificial intelligence, several moral and ethical dilemmas remain unresolved, increasing the risks and challenges of adopting artificial intellengence for business. In business ethics in particular, the rapid evolution of AI technologies, while unleashing opportunities for business, has prompted many important ethical and moral considerations.
While some of these correspond to the AI risks already identified, there are many more:
- Accountability – Who should be accountable? Machines, as such, are not moral agents, and therefore, they cannot be held responsible for their actions. If an AI system violates ethical values, who is responsible for the outcome of that decision-making process? The issue of accountability is further compounded by AI development being mostly outsourced by businesses rather than developed in-house.
- Accuracy, bias, privacy and inequality – Businesses utilising artificial intelligence systems need to make sure that those systems produce precise, reliable and correct results. Since AI learns from data provided by humans, businesses need to have in place appropriate mechanisms to ensure that human biases and prejudices are not mirrored in the AI system’s decision protocols.
- Fair distribution of wealth – Fairness requires businesses to consider the broader impact of AI evolution and development. As AI systems can perform tasks, previously undertaken by humans, more efficiently and reliably, workplace changes will be inevitable. It is, therefore, crucial that businesses pay close attention to how this will impact their employees and customers.
- Integrity – Businesses need to be sure that any AI systems they purchase are used only for their intended purpose. Similarly, when selling an AI system, it is essential to ensure that the use of the AI system by third parties is restricted to the intended purpose.
- Interpretability – As artificial intelligence algorithms become complex, it becomes increasingly difficult to understand how they operate. In a few cases, AI applications have been referred to as a ‘black box’ where even engineers are unable to determine why a machine made a particular decision. Using ‘black box’ algorithms not only creates difficulty in identifying when things start go wrong, it also to makes it difficult to identify responsiblity in case of any ethical lapse or damage.
- Transparency and openness – Due to the lack of skils and experience available, means that global enterprises have spent millions in R&D to develop AI technologies. Many companies which have developed AI algorithms do not allow public scrutiny since the underlying source code (programming) is proprietary. This is crucial since better public knowledge improves trust and prevents unjustified fears.
AI concerns for business
Artificial intelligence is emerging as a defining technology of our age. Mckinsey reports AI is estimated to add $13tr to the global economy by 2030, about 16% of the total global share.
As humans and machines collaborate more closely, businesses need to ensure that artificial intelligence technologies are managed, and subject to, same ethical and moral discipline as any other technology-enabled transformation.
