The opinions expressed by Entrepreneur members are their own.
Last year, many of us pondered the problem of AI bias, carefully described by one of the authors of Coded Bias, the famous Netflix Documentary. Now, with another rise in the popularity of generative AI expected, talk of job replacement is back on track.
Namely, one of the most detailed accounts of how AI can potentially automate (or, as many fear, replace humans in their skilled jobs) comes from Goldman Sachswhich was circulated furiously under many alarmist headlines about 300 million potential jobs worldwide.
In particular, some of the data presented suggests that 18% of work worldwide are likely to be computerized, and the impact on more developed countries may be worse than, for example, on emerging markets.
Oddly enough, the recent boom in generative AI has coincided with several successive waves of layoffs in the online tech industry, which has only made some minor panic in countless online discussions all the more understandable.
Related: 3 principles for building anti-bias AI
However, the report itself suggests that the so-called “exposure to automation” in itself does not in any way imply the elimination or removal of work associated with human participation. More importantly, many non-white-collar occupations are not even subject to negative consequences.
On a broader scale, according to some experts, the ability to use next-generation AI technology will be critical for professionals, instead of being made redundant by solutions like Chat GPT any time soon. Like Ingrid Verschuren, Head of Data Strategy at Dow Jones said“humans are the real ‘machine’ that controls AI.”
Face to face with the reality behind the hype
So, according to Goldman Sachs, almost to 25% of all work can be done by AI completely in the coming years. But what exactly does this mean for a legal specialist, a copywriter, or, for example, a motion designer? To tell the truth, not so much.
A friend of mine who runs a video production studio has been testing AI solutions for imaging for a while, and as it turns out, finding creative inspiration from machine learning algorithms has been quite a tedious journey. The default images are often somewhat generic (and often gloomy for that matter), which is why their design team hasn’t been able to apply much of their newly acquired AI-based assistance.
Meanwhile, in editorial departments, the recent tendency to execute ChatGPT queries on some news personalities and see results that are not entirely truthful has also proven that truthfulness is the weakest point of generative AI.
And given so many false narratives and how easily generative AI tools can be persuaded (e.g. to write content with non-existent facts if provided in a designated request), I highly doubt their legal advice is qualified enough to agree, let alone about replacing even an inexperienced but hungry paralegal with their software equivalent.
Will the future support our fears?
While the current state of generative AI is obviously not as advanced as its founders would like to believe, some labor market forecasts for 2024 may seem overly pessimistic in this regard. There is, of course, a good chance that technology will have a significant impact on our workforce in the coming decade in one way or another. So how can we be ready?
Here are a few key points that entrepreneurs should keep in mind:
Take your time with cutoffs
Whatever niche you work in, the current state of generative AI does not have the skills and competencies to replace any of the skilled people on your team.
More importantly, even as further advancements in AI emerge, you will likely still need your team to manage new software (i.e. explain exactly what needs to be done and then review the result) to get the best results. .
Some of the more prominent examples include code reviews/tweaks, AI script editing, accounting and engineering project reviews, and medical/medical appointment reviews, but the list is almost endless.
On the subject: History has shown what happens to companies that shy away from new technologies, so why are so many people afraid of generative AI?
Check your facts
While we leave the media and celebrities worrying about the possible negative effects of complex deepfakes made possible by the introduction of generative AI upgrades, using ChatGPT or similar tools to search for information remains very difficult.
As learning algorithms evolve, the risks of being completely misled will definitely decrease, but there is a chance that we won’t be able to trust AI generated text/image for the foreseeable future.
While this aspect will continue to be of paramount importance in editorial newsrooms, law firms, and political bureaus, any calculations provided by advanced machine learning algorithms will also require revalidation, at least in selected data cohorts.
Remarkably, the amount of time and operational resources inevitably required to conduct these reviews/reviews actually challenges the conventional wisdom that increased use of AI leads to higher performance with less budget.
Beware of Bias
The first thing we learned when we launched ChatGPT is that its latest “knowledge acquisition is dated 2020-2021”, but more importantly, despite its latest updates, generative AI is still old school, or should we say biased. .
Here are some examples to support my point of view.
I did a simple request by asking ChatGPT to “tell me a story about two people” and got a crappy rom-com about John and Mary. I then ran a short query to draw two people on the beach in the appropriate generative artificial intelligence software, and I got the image of two men (although the scene structure was good, no doubt about it). Presumably, after analyzing my query, the algorithm “decided” that “people” should refer to “male people” in the first place.
For entrepreneurs using generative AI, whether they are in the creative industry or not, this means not only being clear about the risks of AI bias, but also being willing to triple check and then update the intermediate results generated by the software to their inclusion in any additional work product.
Outlook for 2023-2024
In short, whatever misconceptions we have about generative AI at the moment are unlikely to hold true 10 years from now. However, the most reasonable approach to its use remains moderation. Simply put, exaggerating its benefits will certainly be harmful, but over-focusing on its possible ramifications can be just as serious.
To quote Ms. Verschuren of Dow Jones, we humans still have to figure out our future and tweak our machines for better results, no matter how challenging.