AI Risk Considerations

Much to the chagrin of a science fiction author existing in the 1950s, artificial intelligence (AI) is increasingly becoming an integral component in the modern workplace. It can help revolutionize operations, enhance productivity and foster innovation. 

This, however, requires a security-aware workforce to manage the associated risks of such tools. As a reminder, Lindsey Rowe Parker wrote a super fun blog on incorporating AI into marketing practices, but let’s dive a little bit deeper into a few do’s and don’ts of using these AI tools to harness their full benefits, while protecting yourself and your organization. 

Cybersecurity in general is of the upmost priority for our workplaces. (Check out my blog on 10 common traits of phishing emails as another resource on this type of topic.) AI inherently processes data through algorithms and uses this information to make predictions, but this process isn’t stagnant. It operates in a constant feedback loop, where the performance of the model is evaluated and then adjustments are made to improve accuracy each time. This means that information we input into these tools is stored and continually being used as a data point for these AI models. 

Don’t input sensitive company information 

This would include information like: 

• Sensitive personal data: social security numbers, health records, financial information 

• Confidential business information: trade secrets, strategic plans, confidential product designs 

• Customer data: contact information, purchase history 

• Intellectual property: patents, trademarks, copyrighted materials 

• Sensitive financial data: financial statements, banking details, budgetary information 

Do be aware that AI can use incorrect or outdated information 

AI models are trained on data that may not always be accurate. Their performance is dependent on the timeliness and quality of the data they are processing. Cross-check information from multiple resources to ensure the output is correct. 

Don’t lose sight of the importance of human oversight 

While extremely powerful and helpful, AI systems lack the nuanced understanding of human judgement. AI can misinterpret data or overlook contextual factors that a human expert would consider. It is important to maintain a balance between AI and human input to ensure that the results are ethically sound and well-rounded. 

Just as AI is ever-evolving within the workforce, so will our understanding of how to properly utilize these tools for the safest and most correct results. By adopting rigorous security practices and fostering a workforce that is educated and vigilant, embracing AI thoughtfully and proactively can help us reach our goals, while safeguarding each other and the organizations we work for.  

Here are three free AI tools for you to check out:   

1. ChatGPT (OpenAI) – a chatbot that responds to text prompts 

2. DALL·E 2 – an image generator from a text prompt description 

3. Google Colab – a tool to experiment with website code and data analysis 

Be adventurous and forward-thinking, but be safe out there!