Read time: 03'15''
22 September 2023
AI myth-busting: Shedding some (real) light on AI

AI myth-busting: Shedding some (real) light on AI

From letters being signed by tech leaders to halt new AI training to people using ChatGPT to converse on dating apps, AI is certainly starting discussions. In fact, the UN has just lined up a bunch of AI-powered robots in Geneva to carry out “the world’s first human-robot press conference”.

Amongst the stream of content praising AI, lobbying against it and discussing what it is actually capable of, a lot of the truth can get lost or distorted in the online trail. So, to shine some (real) light on the topic, here we take on some common AI myths – and bust them.

AI can think for itself

Despite what many Hollywood blockbusters would increasingly have us believe, AI does not possess independent thought. This is because it’s trained from huge datasets and repositories. Put simply, we control what we want it to know. However, this is where limits can be reached; if we don’t possess the data needed for training, the AI system can not develop its skill set further.

Sometimes, it is even more simple than that, like rules-based chatbots. They only work for the specific scenarios they are trained for and, unlike some AI chatbots, do not learn through interaction.

AI will solve all our problems

With the benefits and risks of AI dominating the news cycle at the moment, it’s difficult to know who to believe. Thanks to AI, breakthroughs in healthcare, education and many other sectors seem to be happening thick and fast. It would be easy to believe AI has the power to fix pretty much every challenge we’re presented with. However, this isn’t true.

While the power and full potential of AI is far from being harnessed, there are still many tasks that it will never be able to accomplish. Similar to its inability to adopt independent thought, AI is based on data that we have provided and therefore can not account for moral or ethical decisions, human emotion or our unique creative spark.

We don’t understand AI

With increasing media coverage, it’s inevitable that tales of scare mongering will arise. A common misconception seems to be that we don’t fully understand how AI works and may be sleepwalking towards disaster.

The algorithms and models used by AI are well understood and documented – they are taught by us after all. In fact, these algorithms have been studied by mathematicians for many years, allowing us to make the advancements we are experiencing regularly today.

AI is only for big tech companies

The myth that AI is only for big tech juggernauts means large companies are missing out on the benefits it can bring. While it’s important to deploy the right AI technology for your business, many advantages can be achieved, including increased productivity, more detailed analysis and quicker decision making.

To remain competitive, large companies should be looking to the future and investing in the right talent. A 2022 study published by the Office for Artificial Intelligence and the Department for Science, Innovation and Technology revealed that over 50,000 people are already employed in AI-related roles, generating £3.7B in gross value added to the British economy.

With this current trajectory, it’s likely to expect the next generation of data experts to be in high demand. It’s up to these companies to empower them and gain an edge moving forward.

Jobs in AI and ML are only for computer nerds

This view has never been more outdated than in the present day. With two-thirds (66%) of large UK businesses revealing they struggle to recruit employees with the skills they need, numerous schemes and partnerships are ramping up to help graduates into data science positions. Incubator startups are also offering significant opportunities to nurture early careers, paving the way for the next generation of data scientists.

With an increased need for jobs to be filled, there is also scope for those outside the tech space to get up to speed on all things AI thanks to the growing number of cheap online courses available. As use of AI grows, expect accessibility to become more commonplace.

AI will take over the world and destroy us all

This is another theme that features heavily in the minds of script writers and film executives. While it’s not completely impossible that AI could be used for perilous means, the right legislation, and ultimately cooperation, can ensure this won’t happen anytime soon.

It’s fitting that the upcoming EU AI Act, poised to regulate artificial intelligence, is set to be published soon. Of course, it’s essential for any new parameters to be policed correctly to have any sort of meaningful impact, but it shows, alongside the growing conversation around ethical AI, that exploitation can be avoided if all corners of the industry work together.

Ed Dixon is the Founder and CEO of Bayezian.