Even if you have a rule against it, your employees probably use ChatGPT for work.
We know this from observation and a few useful statistics.
More than 92% of Fortune 500 companies use ChatGPT. That is according to to OpenAI CEO Sam Altman in 2023.
Higher educated workers and young professionals (64% of users are 18-35) are the largest users of ChatGPT (source).
Never before has a new instrument been introduced so quickly from the bottom up into the workplace. AI is rewriting the way work is done the vast majority of this new hybrid-AI work is done using ChatGPT.
Pretending that your teams don’t use AI is not an option. If you haven’t given them guidelines or training yet, now is the time.
These are powerful tools that can help your organization. But it is important to combine the good with the bad. It’s not as simple as ‘do this, don’t do that’. The way AI impacts employees and businesses is nuanced and requires high-level thinking.
This list of considerations applies to all AI tools. I’m focusing this approach on ChatGPT because that makes it more accessible and easier to understand, since that’s what most people use. Adjust it if necessary for your use.
Companies that best adapt to the impact of AI will consider and respond to all of these considerations.
Start small
AI is a long game that will be with us for some time to come, but quick wins are important here. Quick wins help reduce risk immediately while setting a tone that supports innovation.
Best practice training
This is a great place to start if you haven’t yet organized company-wide training on ChatGPT best practices. A simple one-hour workshop with specific use cases that every department can learn from will help everyone understand what is possible and what to avoid.
Quick policy guidelines
Your employees should have a general idea of what they can and cannot use ChatGPT for when it comes to workplace use. The sooner you get quick policy advice, the better.
It can take months for your organization to become more specific and in-depth Policy and guidance for AI use. In the meantime, you want to make sure you have something that helps the team understand the risks and opportunities.
Determine who is in charge
People across your organization are using AI, from interns to the C-suite.
Start by defining a Center of Excellence (CoE) team that will provide AI leadership, best practices, research, support, and training across the organization.
Your innovation, marketing, sales and legal teams should be represented, as well as someone from leadership.
This provides an enterprise-wide plan and perspective on AI and ChatGPT to guide your organization and prevent inconsistent policies across departments and teams.
That team will begin research to ensure they have a baseline understanding, understand stakeholder opinions, identify pain points, align goals, evaluate tools, and begin identifying model use cases.
Do you have company guidelines for AI use? Click to tweet
Model the best use cases
A recent study of law students who were allowed to use ChatGPT found that they delivered faster, better quality work and enjoyed their work more.
The best way to help your team members get the most out of ChatGPT is to identify where ChatGPT can best help your teams and create best practices around those opportunities.
These examples should be easy to follow. Create clear usage recommendations to mitigate risk and ensure ChatGPT is used to complement human skills, not as a replacement.
For example your Customer service team can use AI tools to quickly gather information and resolve customer issues, but defining a process for them can limit potential problems while maintaining accuracy and high-quality support.
Determine what AI can do well in this process, where it should be avoided and how its quality is controlled.
So besides helping you figure out what to cook for dinner with the 4 random items left in your fridge, what is ChatGPT really good at? You know, for work stuff.
Let’s highlight some of the top use cases where ChatGPT excels:
- Research: Collect and summarize information on various complex topics
- Ideation: Help brainstorm with creative suggestions and alternative perspectives
- Drafting: Produce content outlines and other project scopes
- Edit: Check for spelling, grammar, consistency, and other errors
- Analytics: Extracting key insights and patterns from large sets of data or text
- In summary: condense long documents, reports or meeting notes
- Training: Real-time question and answer support for employee skills development.
- Technical support: resolving common issues and software roadblocks
Not only do these use cases improve productivity and creative output, they also have the potential to take a little “blah” out of the average employee’s workday. At any time you can reduce repetitive and exhausting tasks from your staff will be happier.
Imagine if the car you drive gets a new function every few kilometers you drive on the highway. Ultimately, you need to stop and make sure you still understand how the turn signal works.
Mention the risks
ChatGPT offers a litany of opportunities and improvements for your team members, but we must ignore or shrug off the real risks associated with using AI tools.
1. Bad information
You probably shouldn’t listen to ChatGPT over your doctor just yet. AI tools are getting better every day and hallucinating less, but they are not infallible. We see a steady stream of examples in the media of lawyers, business leaders and politicians spreading false or misleading information.
Does your company have a process in place to check for bad information in all internal and external communications?
2. Bad content
Don’t let ChatGPT do your homework for you.
Over-relying on ChatGPT for copywriting or communications is simply bad practice. Mainly because it’s usually not that good -yet.
ChatGPT typically writes content that is overly verbose, formal, and contains specific turns of phrase that human writers typically don’t include. For example, whenever I see a company blog post with a closing paragraph that begins with “In Conclusion,” I immediately suspect how much of that content was written by a human writer.
This will evolve and improve over time. That makes it even more important to ensure you have processes in place today to ensure quality outcomes for your teams.
Are there expectations for employees to review and rewrite these? content generated by AI.
3. Advanced cyber threats
That unique tone your CEO uses in emails? ChatGPT can probably replicate that quite easily.
As your company gets better at these tools, so do bad actors. Phishing attacks are easier than ever, even for non-native English speakers. Remember, if you have executives with public written, spoken, or video content, anyone could use that data to replicate their voice and cause problems.
With the rise of AI threatsIs your organization increasing attention to cyber threats?
4. Data security challenges
Most leaders probably don’t want their finance team to enter profit and loss data into ChatGPT, saving them 5 minutes of time analyzing some data.
Any information your team members enter into ChatGPT and other AI tools automatically creates data security issues that many companies have not yet grappled with.
A number of risks that your organization must take into account:
- Data Privacy: ChatGPT is trained on data provided during interactions, raising concerns about the unintended exposure of proprietary or confidential information
- Data security: Potential for data breaches or unauthorized access
- Compliance issues: Violations of data protection regulations such as GDPR or HIPAA
- Risk of abuse: Unintentional training of AI models without proper permission
5. AI bias
AI tools, including ChatGPT, can unintentionally reflect and amplify biases in their training data, raising ethical concerns. This can cause ChatGPT to reflect cultural, gender, racial or other biases present in society. It may also contain outdated language or concepts that human editors may better understand.
Be aware of these biases and have a plan to mitigate them in your AI-driven processes.
6. Protect your IP
If your company has a large enough website, AI tools have likely already crawled your content and may be using it in their data sets.
Consider limiting data or content that provides unique value to your business. These could be unique processes, research or community knowledge.
On the other side of the coin, it can be helpful to ensure that AI tools like ChatGPT have access to promotional materials, content marketing, and support resources. These can help your organization surface relevant content for questions from your customers or prospects.
Continuing education
Imagine that the car you are driving is a new function every few kilometers you come to the highway. Ultimately, you need to stop and make sure you still understand how the turn signal works.
This goes beyond technical knowledge. It’s about cultivating an AI-ready mindset. Regular workshops, hands-on training sessions, and ongoing learning opportunities are an opportunity to elevate a team that is not only AI-literate, but also AI-adaptive.
This is a big shift. To avoid becoming overwhelmed, take it slow, but be intentional. And remember: the goal is to augment human intelligence, not replace it with AI. That’s your chance.
