In our modern world of ChatGPT, Alexa and Tesla, Artificial Intelligence (AI) seems to be everywhere. There are lots of ideas floating around about AI; from how it works to its implications for our future. But, along with its growing presence, plenty of misconceptions are made about AI. To try to separate fact from fiction, here are five AI myths debunked.
- AI Will Take Over the World
Science Fiction loves stories of super-intelligent machines plotting world domination. In reality, current AI is nowhere near advanced enough for such feats. This being said, some very clever people, like Bill Gates and Stephen Hawking, have said that there is a possibility that computers may eventually become intelligent enough to start truly thinking for themselves. This is why the companies producing AI, like Google, are employing strict ethical rules, to remove as much bias as possible, and to ensure that AI is being used to our advantage.
In sci-fi, the evil machines have a few things in common; they are autonomous and sentient, and they have the ability to feel some level of empathy or pain; therefore, having a concept of self-preservation. There are all sorts of arguments about consciousness and identity, but the general take away is that for computers to rise up and overthrow human power, they would need to experience wants and pain. These capabilities remain theoretical and simply aren’t viable for our current technology; which leads us to our next myth.
- AI is Approaching Human Intelligence
Terms like ‘neural networks’ and ‘machine learning’ often make it seem like it mimics the human brain. While many AI processes are inspired by human decision-making, AI does not work as an actual human brain would. Neutral networks simply use architecture inspired by human decision-making, allowing computers to compare and analyse information presented to them to recognize fairly complex patterns.
While computers may be able to beat chess masters, or compose music, they do not do so in the same way that humans do. AI simply uses clever algorithms to compare huge amounts of data. For example, it can compare millions of songs to find similarities and create a new melody based on the parameters presented by the program’s operator, i.e., ‘write me a pop song’. They do not understand that sounds make music, nor has it produced a song because it wanted to, it simply uses data to make decisions based on trends.
Furthermore, current AI is not able to feel emotions. It may be able to look at a dataset and find the trends that would denote ‘happy’ or ‘sad, but it cannot actually feel anything. This is way beyond what current technology can achieve, and it is unlikely to be a function that would be implemented by data scientists and AI programmers. This relates back to the ethical rules that AI producers follow. Achieving true sentience or emotional awareness is beyond the scope of current AI, and not an active goal for most developers.
- AI, Machine learning, and Automation are the Same Thing
We have already gone over the basic description of neural networks, but with so much terminology being thrown around, it is hard to actually understand what AI is and isn’t. Let’s add some definitions:
- AI: an umbrella term for technology that enables machines to perform tasks typically requiring human intelligence. It is essentially the science of making a computer or computer program ‘smart’.
- Machine Learning: a subfield of computer science relating to AI. It is a field where computers use datasets and examples to recognise patterns and make decisions and predictions based on these algorithms. Normal computer programs simply follow predefined sets of rules and tasks set by their programmer, rather than actually making its own choices.
- Automation: a process where machines perform tasks based on predefined rules, without any learning or decision-making capabilities. Automation is super useful for businesses, i.e., sending an email to a customer when they purchase a product, or reordering stock when its running low. AI can, however, be used in addition to automation. For example, AI can do market research by analysing trends, and suggest new products to invest in, or to predict how much income a new venture may produce.
- AI Will Take Our Jobs
This is a tricky one. With all technological advances made in human history, jobs have been replaced by machines. The industrial revolution, for example, meant that instead of manually making each individual product, machines could take over these processes and make them on a mass scale. Even with the advent of computers, we no longer needed to handwrite and store paper documents. AI will undoubtedly take over some of the simpler human processes, meaning that some jobs may become obsolete. But, on the flip side, this will actually create new jobs and new ways of working. Processes that were tedious and time consuming for employees can be made simple and quick by employing AI.
In the long run, past technological advances have increased productivity and created new jobs and industries. AI will undoubtedly have the same effect. It is not designed to replace humans, as there are some things that AI simply cannot do in the way that humans can. Integration between AI and a workforce may be challenging, but with proper training, and industry wide standards and policies, it is certainly possible.
- AI is Expensive and Only for Large Businesses
Traditionally, AI has been expensive to implement. It cost around $5 million (£4.7 million) to develop ChatGPT. The reason for these huge expenses is that it costs to gather, cleanse, and store all the data necessary to provide computers the power to process the huge amount of information needed for an AI system to work.
This being said, AI services are becoming far more accessible thanks to the Cloud. They can be accessed and used at low cost, and without any specialist knowledge. This is having the effect of standardising and equalising the industry, enabling enterprises of all sizes to benefit from AI-driven insights and efficiencies that were once exclusive to large corporations.
—
AI continues to evolve, and while it’s reshaping industries and redefining possibilities, it’s important to separate fact from myth. By understanding AI’s capabilities and limitations, we can better harness its potential to improve our lives and work.
All of this being said, AI also presents significant challenges. One such concern is its use of existing data, particularly in the creation of AI-generated ‘artwork’. Many artists and online communities strongly oppose this practice, as AI models often learn from artwork without the original creators’ consent. Additionally, the use of AI-generated media has been linked to issues involving false advertising and deepfakes, raising ethical and security concerns.
These risks highlight the potential dangers of AI becoming commonplace. Companies and businesses looking to integrate AI into their operations must do so with a clear understanding of what their AI is being used for, and putting controls in place to stop theft, deep-fakes and misleading information or advertising. While AI is a powerful tool that can improve efficiency and innovation, it is vital that it is developed and deployed ethically to ensure it serves society fairly and responsibly.
