Risk Trends 2025 and Beyond: Artificial Intelligence Ethical Dilemmas

Can you say with confidence that AI will be more of an opportunity than a risk?

The Canadian artificial intelligence (AI) market is booming. Current projections have it reaching US$4.13 billion in 2024 and growing to US$18.5 billion by 2030, according to Statista. AI is transforming businesses, making them smarter and more efficient. But AI brings a whole host of ethical dilemmas that organizations must navigate.

One of the biggest concerns is bias and discrimination. AI systems are often programmed with distinct biases. For example, an AI system might favour one group over another simply because the programmer influenced the code to prioritize certain options, or the training data was biased. 

To combat this, businesses must ensure that AI is programmed and trained on diverse and representative data sets. Also, the AI may need complementary technology to help it make better decisions, such as facial recognition scanners. Regular audits and testing can also help detect and correct these unwanted biases.

Privacy is a significant issue with AI’s data collection and analysis capabilities, which can threaten individual privacy by identifying people and predicting personal attributes. This can lead to misuse of personal information, such as inadequately protected health data from a smart watch. Strong governance practices are essential to protect privacy and ensure ethical data use. 

Be prepared, not sorry

Misinformation and deepfakes are also raising concerns. AI can create convincing content to spread misinformation or commit fraud. Organizations need to develop tools to detect and counteract deepfakes and fraud threats, ensuring the integrity of their information and their systems. 

There is also a need to educate individuals to be skeptical about who is establishing contact. For example, if someone suspects the person calling is not who they claim to be, they should ask a complex question that only the genuine person would know the answer to. 

AI systems are not immune to attacks. Adversarial attacks can manipulate AI inputs to produce incorrect outputs, leading to security breaches or bad decisions. Implementing robust security measures, like adversarial training and continuous monitoring, can help keep AI systems safe and functioning as intended.

The rise of AI-driven surveillance systems raises concerns about mass surveillance and its impact on civil liberties. Organizations must ensure that their surveillance practices are transparent, ethical, and protective of individual rights to maintain public trust. 

Lastly, AI can be used to enhance cyber attacks, create persuasive misinformation campaigns, and automate spam and phishing. To protect against these threats, businesses need to stay vigilant and proactive in developing controls and countermeasures.

Why stop there? Here are other risks to consider:

• Misuse of autonomous drones and weapons
• Job displacement when technology replaces manual labour
• Lack of explainability due to user ignorance of code biases
• Everyday technologies that, unbeknownst to individuals, copy, store and use data from their laptop, smartphone, or watch

Questions to consider:

• How are trends in artificial intelligence impacting your organization and what are you doing to get ahead of it? 
• Do you have an inventory of all the ways you’re using AI? Has your organization created a policy identifying how to use it correctly?
• Have you asked your software and hardware providers how they intend to use AI? 
• Can you turn off unacceptable AI use?