Many organizations are at a critical crossroads in the race toward AI-driven automation. The allure of AI is irreristible. It promises greater efficiency, higher productivity, and streamlined operations. But, as with any technological revolution, the question is whether it’s a necessity or just a trend.
We talked to Shawn David, an AI expert, about why companies must adapt to AI automation to survive and boost their productivity. We also discussed the potential risks AI technologies entail.
Shawn David has degrees in computer science and organizational leadership. He found his passion in automation while working at a marketing agency. He now runs "Automate to Win," educating entrepreneurs on using AI to enhance productivity and efficiency.
NordLayer: Shawn, generative AI tools like ChatGPT and Midjourney are now on everybody’s radar, and some organizations have already successfully deployed them. What are the benefits of using AI for businesses?
Shawn David: There are many benefits, but first of all, AI greatly improves business efficiency and productivity, particularly in automation. Employees can swiftly spot patterns and segments in large datasets, saving them from the manual analysis of thousands of lines of spreadsheets. This speed and accuracy can help combat fatigue and reduce errors. For example, creating a presentation deck can take 20 minutes instead of five hours.
By automating manual and repetitive processes, businesses can focus on understanding their workflows and automating the right processes. The essence is in deploying automation identifying and optimizing processes for reliable outputs. It’s the bread and butter of AI-driven automation, which, done properly, helps businesses achieve greater efficiency, higher productivity, and more reliable results.
NordLayer: You said repetitive jobs would be automated. And interestingly, Gartner predicts that 40% of such tasks we do now will be automated by 2030. What are the business implications of adopting AI?
Shawn David: I can highlight an example from an agency’s context. This agency conducts more than 2,000 A/B tests annually across various clients. Now, imagine streamlining this and automating the mechanism. Whatever time and workforce you save can be entirely freed up. Also, assuming the agency’s billable rate is $300 per hour and the average time spent on this particular task is almost five hours, the annual gain amounts to $2.7 million. What’s remarkable about AI is that this isn’t just a cost-saver. It’s a strategic asset that fuels growth, boosts profits, and eliminates repetitive tasks.
NordLayer: Indeed, it’s remarkable, but the question is how to start with AI in business. Can you explain your three-step strategy for organizations looking to leverage AI?
Shawn David: Think of the first step as testing things out. You must step beyond your usual methods and rethink how you use AI. Even I, with two decades of systems engineering, had to adjust my thinking after about six months of working with it.
Consider three generative AI models: Claude, Bard, and ChatGPT. Think of them as three different ovens, each baking the same cake but with unique qualities and results. You are the chef here.
And now, the idea is to use AI to create a recipe for optimizing your business. Start with the end goal. AI will show you how to use the ingredients (your data and processes) to reach that outcome. Remember that you can hire a professional if you’re unsure how to use AI effectively for your specific needs.
NordLayer: First, you pinpoint your business needs with AI. What's the next step?
Shawn David: You use AI with your data in the second step. There are two paths here: public Large Language Models (LLMs) like Bard and Claude are great for answering questions you already grasp. Imagine the result you desire. For instance, getting insights from 12,000 PDFs. The beauty is LLMs understand what you ask.
In the third step, you apply AI to your business processes. Now, avoid the DIY route. There are professionals who can handle this for you. Also, building a customized AI tool will help you secure your data by providing a more advanced and proactive approach and preventing unauthorized access to sensitive information.
NordLayer: Let's explore a practical scenario. Say I'm a healthcare business with around 60 employees, and I'm aiming to develop a secure AI tool.
Shawn David: Absolutely, building a safe AI tool involves a step called vectorization, which is like having a super-smart assistant. It studies content, creating connections and context within data. For example, when we think of "jaguar” and instantly associate it with "Kitty Cat" or "English car."
If you have sensitive data, like personal information, create an isolated database, keeping it safe from outside networks. Then, use your natural language processing on this database without directly linking to an AI model.
As you invest in vectorization, you eliminate complex translations between humans and machines. Encoded vectors carry meaning, easily understood by AI. This ensures clear communication, fostering secure and context-aware AI tools.
In summary, for goals like healthcare data privacy, vectorization empowers secure, efficient, and direct interactions with AI systems.
NordLayer: Certainly, understanding risks is crucial. Are there downsides to deploying a customized AI tool for businesses?
Shawn David: One concern is not fully knowing how it works and blindly trusting its results. There's a risk when you get output without clarity on the process. Let's look at an example. Imagine you purchase a natural language processing system for your company. You feed in data, which tells you 51 to 85-year-old females have a 97% conversion rate. But it might hide the idea that grandmothers are buying kids' presents, leading to wrong business decisions based on misleading info.
To avoid this, you need to understand the AI’s mechanism. Talk to engineers, ask about biases, and get answers. If not, bring experts or consider adopting open-source solutions in a safe, air-gapped (disconnected) environment. This keeps data secure. It’s like driving a tractor-trailer if you don't know how you’re stuck. Unlike in your first step, trial and error won't help here. You need a solid understanding to navigate potential challenges. I advise talking to AI tool developers with the insights for smooth and secure use.
NordLayer: How do the risks you highlighted align with the ever-evolving landscape of cyber threats, and what challenges arise at the crossroads of AI and cybersecurity?
Shawn David: The emergence of AI-driven attacks has elevated the game beyond the usual culprits like software bugs or human errors. These attacks are woven into the very fabric of algorithms, and fixing them is now a complex puzzle.
Interestingly, despite their sophistication, many AI attacks exploit the same traditional vulnerabilities, such as weak passwords, unpatched software, or social engineering. In simpler terms, AI introduces new threats and uses age-old weaknesses. Strengthening cybersecurity defenses with established measures like strong passwords and multi-factor authentication can indeed create a formidable barrier.
NordLayer: Your insights on AI are intriguing. Shifting gears to the blend of cybersecurity and AI, especially in light of the growing concern surrounding deepfake and video content, what narrative do you see taking shape?
Shawn David: The proliferation of deepfake audio and video content undoubtedly demands our attention. With the rise of the LLMs and the whisper-1 audio-to-text and then training, you can quickly recreate someone's likeness in real-time audio.
AI can mimic words you never uttered, prompting a quest for alternative validation methods. A human-based certification of any digital communication, perhaps? If AI can 100% replicate my voice and appearance during a Zoom call or run my LinkedIn posts, that's scary.
NordLayer: So what, in your opinion, is the smartest way to ensure data safety within the realm of AI?
Shawn David: The data that AI feeds on can be twisted around for bad purposes in totally new and unexpected ways. This means we must change how we gather, keep, and use that data when dealing with AI.
Think of it like building your own AI tools in-house. It’s just like creating a strong fortress to safeguard your valuable assets. As I said earlier, adopting open-source solutions in a safe, disconnected environment keeps data secure. My advice is to use a thoughtful approach to AI and cybersecurity. This will help make sure a business remains resilient.
NordLayer: That’s an insightful perspective on cyber threats in the context of AI. Now, let’s delve into AI-powered decision-making. Can you explain what it means?
Shawn David: According to Gartner, by 2025, 95% of decisions involving data will be at least partially automated. AI will improve the speed and accuracy of decisions in three ways.
First, we’ll have human-based decisions, such as medical diagnoses, where machines assist with visualization, but humans make the final call based on ethics, bias, logic, skills, and emotions.
Second, we’ll have hybrid decisions, like in financial investment, where the machine suggests, but the human decides. AI will provide recommendations and analytics for human validation.
Finally, we’ll have full decision automation, as in choosing the next best action for a digital order, where the machine decides using predictions and forecasts. Managing risks is key here, which might involve setting guardrails or keeping a human in the loop.
NordLayer: Now, let’s talk about the impact of AI across various industries. How can different sectors benefit from AI integration?
Shawn David: Absolutely, the potential applications are vast. Industries characterized by routine human interaction or manual tasks are ripe for transformation. Consider roles involving data transposition, low-level content creation, or entry-level graphic design. For instance, I'm developing a system that can replace low-paying content creation gigs on platforms like Upwork and Fiverr. Users can train an AI using their own social media posts, and it will generate content in their style, which can then be refined further. This kind of symbiotic relationship between humans and AI can lead to a more efficient workflow.
Consider facilities like the Cleveland Clinic and the Mayo Clinic, which already utilize AI to monitor patients round-the-clock. Vital signs, such as heart and breath rates, are tracked, alerting medical staff to changes. AI aids in medication dispensing and even assists in medical diagnosis, analyzing patterns in cells that the human eye might miss. Radiology and surgery benefit from AI's unmatched precision, akin to a drug-sniffing dog that never fails to detect. The potential of AI in healthcare is profound, touching everything from immediate care to complex diagnostics.
NordLayer: It's interesting how AI's influence varies across industries. What about the IT industry? Will it also see a significant impact?
Shawn David: Absolutely, even in the IT industry, there are areas that AI can revolutionize. For instance, AI bots can easily handle routine tasks like answering basic customer queries about platforms like GoDaddy or providing guidance on Google or Facebook ad setups. AI can watch and understand training videos, making manual checks unnecessary. This kind of automation can reshape entire sectors overnight. However, it's important to note that AI adoption in IT is all about efficiency and profit. Open AI's training data includes questions from users, not as an altruistic gesture, but to improve AI's performance.
Regarding industries with less potential AI impact, those heavily reliant on high-level creativity or personalized concierge services might not experience significant change. For example, luxury services that thrive on human touch and bespoke experiences may remain less influenced by AI's reach.
NordLayer: Finally, can we discuss risks for businesses regarding AI?
Shawn David: One of the biggest business risks is falling behind the AI race and losing competitiveness. If you approach automation authoritatively without involving the workforce, it can lead to resistance and chaos. The key is to view AI as a collaborator, like a helpful robot. If people grasp this concept, we’d be in a better place. However, the rush to adopt AI while disregarding ethical concerns can lead to unintended consequences. It’s crucial to strike a balance between progress and responsibility.
NordLayer: Thank you very much for this insightful conversation.
Shawn David: My pleasure.
Before diving into the AI world, consider securing your business’s digital journey. Learn how NordLayer can help you strengthen your defenses.
This text has been generated by a human
Joanna Krysińska
Senior Copywriter
A writer, tech enthusiast, dog walker, and amateur pastry chef, Joanna grew up in a family of engineers and mathematicians, so a techy mind is in her genes. She loves making complex tech topics less complex and digestible. She also has a keen interest in the mechanics of cybercrime.
Subscribe to our blog updates for in-depth perspectives on cybersecurity.
