10 Things Every Business Leader Needs to Know About AI in 2025
By Ben Jones, CEO of Data Literacy and author of AI Literacy Fundamentals
Whether you run a business or lead corporate teams, there’s no avoiding artificial intelligence (AI). It’s everywhere in 2025, along with hype, fear, and confusion. Here are ten things every business leader should know about AI to navigate today’s choppy waters.
1. You’ve already been using AI for a long time
While ChatGPT’s 2022 launch dramatically increased usage of generative AI, the fact is we’ve been using AI for decades. The field of AI has a long history, stretching back to the 1950s, when researchers began creating machines that could perform tasks typically requiring human intelligence.
What we call “AI” continues evolving, but you’ll find more than just chatbots underneath its expanding umbrella. You’ll also find computer vision, machine learning, and robotics, to name a few of its major subfields. Whenever you’ve used facial recognition, spam filters, recommendation engines, or fraud detection systems, you’ve been using AI.
2. Your employees use Generative AI with or without your approval
In 2023, stories emerged of companies like Samsung that banned the use of generative AI on company devices after employees shared sensitive content with ChatGPT. Since early versions of AI chatbots trained their models on users’ chat sessions, other companies followed suit with their own bans.
Studies found, however, that over half of global workers used generative AI without their employer’s formal approval. AI chatbots are evolving, and most now allow users to opt out of giving chat contents for model training. Additionally, many enterprises have deployed their own internal chatbots with security features. Nevertheless, many company-wide bans remain in place. Employees are ignoring them.
3. Deploying AI for the sake of it won’t help
Here’s what doesn’t work in 2025 (or ever): rushing to implement new tools because that’s what everyone’s doing. That’s not leading, that’s following. Great business leaders find ways to apply technologies strategically rather than blindly following trends.
What problem are you trying to solve? What opportunity are you pursuing? AI can help reduce costs, increase efficiencies, and open up new markets, but it’s not a panacea. Generative AI, in particular, is great at some tasks and terrible at others, resulting in a “jagged frontier” that’s difficult to navigate.
4. You need an AI strategy that you revisit quarterly
And this jagged frontier is constantly shifting. Not a week goes by without a major release, announcement, or launch. It’s hard to keep up with it all.
Given this frenetic pace, you should develop a strategic plan for AI implementation. Revisit it quarterly with your team, and encourage discussion about relevant advancements so that you can adjust your tactics quickly if necessary.
5. Deploying AI comes with risks and downsides
As powerful as AI is, it’s far from perfect. It will propagate the flaws and biases in its training data. AI chatbots hallucinate, providing incorrect or nonsensical outputs.
There are ongoing legal battles over copyright infringement, ethical concerns about the exploitation of low-wage workers who flag disturbing content, and environmental costs of data centers for training and operating state-of-the-art systems. Leaders should weigh these risks and downsides. Ignoring them won’t make them go away.
6. It’s crucial that your employees develop AI literacy
In their inaugural “Skills on the Rise” report, LinkedIn ranked AI literacy as the fastest growing skill of 2025. AI literacy is “the ability to recognize, grasp, use, and critically assess artificial intelligence technologies and their impacts.”
AI literacy is so important, it’s now legally mandated in Europe. Beginning in August 2025, the European Union will begin enforcing Article 4 of the EU AI Act, requiring businesses to ensure “a sufficient level of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf.”
Whether or not this law applies to your organization, make sure your employees develop AI literacy. You can’t afford to fall behind in this area.
7. AI training needs to be role-specific and hands-on
Don’t roll out a one-size-fits-all training program and then force everyone to take it. Such mandatory corporate training rarely does any good.
Instead, define employee personas based on tasks and responsibilities, and then create custom learning pathways for each group, complete with hands-on, practical workshops. If you want them to learn about AI, they have to try it.
8. You can customize AI using your company data
Today’s foundational models weren’t trained on your company’s documents, databases, or source code. It’s possible, however, to add proprietary content to knowledge bases that models can access using an approach called Retrieval Augmented Generation (RAG).
Such customized systems are able to respond to prompts from your employees in a more tailored manner, making them more relevant than generic systems. And with platforms like ChatGPT Team and Claude for Enterprise, secure and cost effective systems are available even for small and medium sized businesses.
9. You need to verify AI-generated outputs carefully
AI chatbots commonly provide confident-sounding outputs, but that doesn’t mean the outputs are correct. You need to cross-check and fact-check every single output that any AI system provides. I can’t stress this enough.
While customized AI models are more likely than generic models to provide useful responses, they’re still susceptible to inaccuracies and hallucinations. Generative AI in 2025 is incredibly powerful, but it’s still quite error-prone.
10. You can measure the AI’s impact on your business
Any good business leader sets goals and objectives, and implements metrics to track progress. Implementing AI is no different. Are you hoping to achieve sales increases, cost reductions, quality improvements, process efficiencies, or some combination of these aims?
Be specific about the measurable outcome you’re looking for, and then hold yourself accountable by publishing the metrics for all to see. This is Change Management 101.
Conclusion
AI is evolving quickly, which is both a challenge and an opportunity. The savvy business leader pays close attention to what’s happening, discerning between helpful and harmful aspects. If you foster a balanced and healthy dialogue about AI, you’ll have a much better chance of succeeding.
(note: a version of this article was previously published to the Forbes Business Council)
10 Things Every Business Leader Needs to Know About AI in 2025
Whether you run a business or lead corporate teams, there’s no avoiding artificial intelligence (AI). It’s everywhere in 2025, along with hype, fear, and confusion. Here are ten things every business leader should know about AI to navigate today’s choppy waters.
1. You’ve already been using AI for a long time
While ChatGPT’s 2022 launch dramatically increased usage of generative AI, the fact is we’ve been using AI for decades. The field of AI has a long history, stretching back to the 1950s, when researchers began creating machines that could perform tasks typically requiring human intelligence.
What we call “AI” continues evolving, but you’ll find more than just chatbots underneath its expanding umbrella. You’ll also find computer vision, machine learning, and robotics, to name a few of its major subfields. Whenever you’ve used facial recognition, spam filters, recommendation engines, or fraud detection systems, you’ve been using AI.
2. Your employees use Generative AI with or without your approval
In 2023, stories emerged of companies like Samsung that banned the use of generative AI on company devices after employees shared sensitive content with ChatGPT. Since early versions of AI chatbots trained their models on users’ chat sessions, other companies followed suit with their own bans.
Studies found, however, that over half of global workers used generative AI without their employer’s formal approval. AI chatbots are evolving, and most now allow users to opt out of giving chat contents for model training. Additionally, many enterprises have deployed their own internal chatbots with security features. Nevertheless, many company-wide bans remain in place. Employees are ignoring them.
3. Deploying AI for the sake of it won’t help
Here’s what doesn’t work in 2025 (or ever): rushing to implement new tools because that’s what everyone’s doing. That’s not leading, that’s following. Great business leaders find ways to apply technologies strategically rather than blindly following trends.
What problem are you trying to solve? What opportunity are you pursuing? AI can help reduce costs, increase efficiencies, and open up new markets, but it’s not a panacea. Generative AI, in particular, is great at some tasks and terrible at others, resulting in a “jagged frontier” that’s difficult to navigate.
4. You need an AI strategy that you revisit quarterly
And this jagged frontier is constantly shifting. Not a week goes by without a major release, announcement, or launch. It’s hard to keep up with it all.
Given this frenetic pace, you should develop a strategic plan for AI implementation. Revisit it quarterly with your team, and encourage discussion about relevant advancements so that you can adjust your tactics quickly if necessary.
5. Deploying AI comes with risks and downsides
As powerful as AI is, it’s far from perfect. It will propagate the flaws and biases in its training data. AI chatbots hallucinate, providing incorrect or nonsensical outputs.
There are ongoing legal battles over copyright infringement, ethical concerns about the exploitation of low-wage workers who flag disturbing content, and environmental costs of data centers for training and operating state-of-the-art systems. Leaders should weigh these risks and downsides. Ignoring them won’t make them go away.
6. It’s crucial that your employees develop AI literacy
In their inaugural “Skills on the Rise” report, LinkedIn ranked AI literacy as the fastest growing skill of 2025. AI literacy is “the ability to recognize, grasp, use, and critically assess artificial intelligence technologies and their impacts.”
AI literacy is so important, it’s now legally mandated in Europe. Beginning in August 2025, the European Union will begin enforcing Article 4 of the EU AI Act, requiring businesses to ensure “a sufficient level of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf.”
Whether or not this law applies to your organization, make sure your employees develop AI literacy. You can’t afford to fall behind in this area.
7. AI training needs to be role-specific and hands-on
Don’t roll out a one-size-fits-all training program and then force everyone to take it. Such mandatory corporate training rarely does any good.
Instead, define employee personas based on tasks and responsibilities, and then create custom learning pathways for each group, complete with hands-on, practical workshops. If you want them to learn about AI, they have to try it.
8. You can customize AI using your company data
Today’s foundational models weren’t trained on your company’s documents, databases, or source code. It’s possible, however, to add proprietary content to knowledge bases that models can access using an approach called Retrieval Augmented Generation (RAG).
Such customized systems are able to respond to prompts from your employees in a more tailored manner, making them more relevant than generic systems. And with platforms like ChatGPT Team and Claude for Enterprise, secure and cost effective systems are available even for small and medium sized businesses.
9. You need to verify AI-generated outputs carefully
AI chatbots commonly provide confident-sounding outputs, but that doesn’t mean the outputs are correct. You need to cross-check and fact-check every single output that any AI system provides. I can’t stress this enough.
While customized AI models are more likely than generic models to provide useful responses, they’re still susceptible to inaccuracies and hallucinations. Generative AI in 2025 is incredibly powerful, but it’s still quite error-prone.
10. You can measure the AI’s impact on your business
Any good business leader sets goals and objectives, and implements metrics to track progress. Implementing AI is no different. Are you hoping to achieve sales increases, cost reductions, quality improvements, process efficiencies, or some combination of these aims?
Be specific about the measurable outcome you’re looking for, and then hold yourself accountable by publishing the metrics for all to see. This is Change Management 101.
Conclusion
AI is evolving quickly, which is both a challenge and an opportunity. The savvy business leader pays close attention to what’s happening, discerning between helpful and harmful aspects. If you foster a balanced and healthy dialogue about AI, you’ll have a much better chance of succeeding.
(note: a version of this article was previously published to the Forbes Business Council)
