Hype or AI Revolution? Or maybe it’s both.

With all the hype that AI brings, it’s easy to mistake it as 2023’s philosopher stone. Like a cure-all potion or a lamp with a magic-granting genie, innumerable companies tried (and still are trying) to plaster AI features to their products, in what seems to be a desperate attempt to capture an inch more of their market share.

On the outside though, it’s starting to look like a bad case of FoMO (Fear of Missing Out) instead of a deliberate attempt to add product value.

The litmus test of whether adding AI to a product is a good idea is by asking yourself these three questions in this particular order:

• Do we have the infrastructure to support AI development?
• Can we build AI-powered features fast enough given the current tech crunch?
• Am I adding AI to the product in such a way that it improves performance?

If you’ve answered ‘No’, then consider if adding AI is a good value-add or something that’s best kept as a marketing event.

AI is not Magic. It’s Math: equations and probabilities.

Despite the hype, most AI products are simply probability calculators at their cores. Many problems that they aim to solve revolve around dealing with uncertainty.

Consider ChatGPT. When someone asks it a question, it uses the data fed to it, run them through its algorithm, and returns the most probable correct answer. This can be done because it follows a set of mathematical formulas specifically designed to provide the best answer.

Depending on what it is meant to do, it can also be trained by adjusting coefficients in such a way that it can return the outcomes we expect it to provide. For many neural networks like ChatGPT, the goal is to return close to accurate answers while remaining lifelike and responsive.

While chatting with an AI-powered chatbot may seem like talking to a human, it is merely a machine attempting to replicate human responses by using crafty models and humongous amounts of data.

In a sense, AI is a continuous optimization problem to give the best probable output using rules expressed in mathematical formations.

Again, AI is not magic. It’s Math.

The AI mistake threshold: Hiring an Expert Copywriter

There are a lot of conversations boldly asserting that AI will ‘replace all jobs’. And while it’s too early to say for sure, there are enough reasons to believe that the technology will inevitably arrive there.

The questions now are when and how much. One thing is for sure, it’s not going to be the apocalyptic kind that many are hyping it to be.

Broadly speaking, what AI is attempting to replicate is the robust decision-making skills of real people. But it’s clearly not there yet. It can only operate properly within the parameters it was trained with.

Adding an unknown element can cause gross inaccuracies - something real people are adept at adjusting to. No matter how much embellished AI stories can read like, there’s still no substitute for a real and trained human person. For now, at least.

Humans excel at making very accurate decisions. We spend years to decades studying, training, and working to reach a level of expertise that allows us to do things better than other people. When we do, we get to charge more for our ‘experience’ because we’re efficient and effective compared to all other options.

In general, businesses can leverage expertise because they can steer organizational operations away from resource traps - like learning from costly mistakes or choosing tools and campaigns that are delivering suboptimal results. There are also industries that will try to minimize inaccuracies as much as possible either due to regulation (medicine) or because the very nature of the industry does not allow it to make mistakes (law, banks).

When AI is added to the equation, it replaces the high-ticket tasks humans can do but are so resource-intensive or repetitive that it is not cost-effective to train and/or hire an expert, when there’s a certain level of acceptable threshold for it to make mistakes.

For example, copywriting. While it’s important to know that different words evoke different meanings when they’re alone or strung along in a sentence, for the vast majority of small to medium businesses, a 5% increase in effectiveness because of word choice does not justify hiring an expert copywriter that could have worked for Google or Apple instead. To a struggling mom-and-pop shop, a 5% increase is negligible. For a multi-billion dollar company like Amazon, a 5% increase in profit margin spells the difference between potential layoffs of thousands of people or company-wide bonuses.

The same is true for other industries. Take video editing for instance. There are a lot of amazing AI tools like CapCut that could bridge the skill gap from a beginner video editor to a more experienced one, as these tools are able to provide higher quality outputs with less learning curve — just like how the easier Canva became the entry-level tool for social media managers, replacing Photoshop which could do a lot more, but takes time to learn.

AI allows you to scale by freeing up important resources for your team - like time and talent - at the cost of making mistakes. Depending on the industry, users, and market, that threshold will look different.

As an AI Product Manager, it’s your job to determine if the trade-off is worth the risk. You have to find the sweet spot between driving value and how much ‘inaccuracies’ are acceptable. Since AI relies on probabilities instead of expertise, it will inevitably be wrong. You have to wade through the murky water and assess how much AI can do wrong before it becomes a faulty tool that is not worth buying.

‍

What’s the verdict: Is AI worth the hype?

Yes, it’s worth the hype. For the most part, AI is doing great strides in liberating humans from laborious and resource-intensive tasks. It has a wide array of applications that could revolutionize the way we do things. We can become more effective with our time, allowing us to allocate resources to more important initiatives.

But it has its current limitations. Technical constraints are legitimate criticisms. And it will take time to be the epoch-defining technology it has hyped itself to be.

The AI field is an exciting place to be in. However, no one is just going to let you come in without qualifications precisely because it’s a highly sought-after industry. That's especially true when the market is saturated with talent.

If you want to shift to tech, especially from a non-tech background, you need to either have good domain knowledge, lots of transferrable skills, or have something to bridge the skillset you have currently and the requirements of the roles that you are looking for.

One of those places is Co.Lab, where you can get real-life work experience by building a live product of your own with a cross-functional team in an agile environment. All with a supportive community of aspiring techies like you.

Because we believe that —
You Belong in Tech ✨

—

Are you an aspiring Product Manager, Product Designer, or Software Developer? The Co.Lab program is the perfect place to gain real-world, cross-functional experience that you wouldn’t get anywhere else. Follow us on on Instagram, Twitter, and LinkedIn for the latest updates.