Reducing AI-related risks through effective internal communications

Key Takeaways:

• Many businesses are exploring the potential of artificial intelligence, but few have formal governance in place.
• AI policies are essential for risk mitigation, whether they guide its use or prohibit it.
• Simply having AI policies, however, is not enough. Companies must take a more proactive approach to training and implement ongoing communications initiatives.
• You can download a free self-assessment to identify weaknesses in your communications strategy by clicking here.

The Rise and Risks of AI

Even sectors that have historically been slow to adopt new technologies are exploring the opportunities that generative artificial intelligence presents for enhancing efficiency and its potential cost-savings. Unfortunately, an alarming number of companies are overlooking the need for internal communications initiatives surrounding it.

This oversight is not just problematic for those on the cutting edge; it also spells potential trouble for businesses that would prefer if their employees didn’t use AI.

A Concerning Lack of Governance

According to 思科 's 2024 Privacy Benchmark Study, approximately 27% of companies have banned the use of generative AI, while a survey conducted by ISACA found that 28% reported expressly permitting it. Of those that allow it, only 10% reported having a comprehensive framework in place. One in four said no policy exists and there were no plans to create one.

Without explicit policies around artificial intelligence—whether to guide its use or to limit it—companies leave themselves vulnerable to unpredictable, perhaps unmanageable, risks.

The Importance of Internal AI Policies for Risk Mitigation

A lack of framework and direction can lead to misuse, which can have significant consequences. Let’s take a look at a few examples:

1. Inaccurate Data and Misguided Decisions

Scenario: Intent on delivering comprehensive market insights under a tight deadline, an employee uses an AI tool to conduct research on the web and develop strategic recommendations. Trusting the AI’s accuracy, they neglect to conduct due diligence by not fact-checking the information before presenting it.

Risk: Inaccurate or misleading data being used in decision-making.

Potential Outcomes:

• Strategic Missteps: Critical decisions based on flawed data could result in failed strategies and direct financial losses
• Loss of Credibility: Exposure of inaccuracies could damage credibility with key stakeholders, eroding trust and potentially impacting partnerships and investor relations
• Resource Waste: Significant resources may be misallocated, wasting time, money and effort on unviable initiatives

2. Sensitive Data Exposure

Scenario: Under pressure to quickly analyze customer behavior trends, an employee decides to use a third-party AI tool to sift through vast amounts of sensitive customer data. Believing the AI would expedite the analysis, they bypass internal security checks and do not seek approval for the tool’s use.

Risk: Unauthorized sharing of sensitive data with an unsecured AI tool.

Potential Outcomes:

• Data Breach: Possible exposure of customer data, risking identity theft and financial liabilities for the company
• Penalties and Litigation: Violations of data protection laws, resulting in significant fines and legal issues
• Reputation Damage: Loss of customer trust and loyalty, leading to lasting damage to the company’s public image

3. Compromised Content Integrity

Scenario: In an effort to keep up with content demands, a marketing employee uses an AI tool to automate the creation of promotional materials. The content is published without thorough review to ensure it aligns with the brand’s voice and values.

Risk: Content may not align with the brand, may contain errors or seem disjointed.

Potential Outcomes:

• Misrepresentation: Unvetted content could misconstrue the message, leading to a public relations crises
• Customer Trust Decline: Inaccuracies or inappropriate content can alienate customers, reducing engagement and loyalty.
• Operational Inefficiency: Time and resources might need to be redirected to address and rectify the fallout from the content errors, affecting overall marketing productivity.

AI policies alone are not enough

According to Salesforce , over a quarter (28%) of workers use generative AI at work, with approximately 55% leveraging tools that have not been approved by their employer. More concerning is that nearly 40% admit to using ones explicitly banned by their organization, underscoring the need to do more than just implement new policies.

Widespread non-compliance doesn't necessarily stem from a place of defiance. Often, it is because of poor communication and a lack of understanding. Salesforce found that about 70% of employees haven’t received training regarding the use of AI ethically, or in general. As a result, many fail to grasp the full extent of the liabilities involved or how their actions could potentially contribute.

This can be addressed through specialized training and communications initiatives that convey the seriousness of these risks and their consequences in a way that resonates with employees. To maximize impact, these educational efforts should be tailored to different roles within your organization. Ongoing engagement helps ensure the information stays top of mind.

Additionally, creating an open environment where employees can ask questions and discuss their concerns about AI use can lead to better understanding and adherence to policies.

Are you effectively communicating the risks associated with AI misuse? Click here to download our free, simple self-assessment.

If you’re ready to take a proactive approach to risk mitigation, we can help. Get in touch to schedule a free consultation! Send us a message here or by emailing [email protected].

Looking for more expert insights? Subscribe to our newsletter.