What is bias?

Bias is everywhere, it is built into our genetic code, leaning towards one thing more than another. Bias is the simple idea of prejudice against something and in this case, the something being the whole idea of AI and possibly what AI even stands for. Many people have a very strong pre-conceived opinion on AI and that goes both ways, whether that be good or bad?  Much of this reason of bias is due to mistrust and misinformation, most of this coming from many journals and tabloids that give alternating stories depending on their usual audience.  Older generations tend to bias towards ‘not liking it’ and younger generations the opposite, as a study finds that older people are quicker to reject new technologies (Lancaster, 2018) even with AI being at the forefront of healthcare technological advancements with machine learning being used to predict early signs of many diseases and illnesses such as cancer. (Thomas Davenport, 2019)

How will bias effect AI?

There are many reasons why such bias will prevent, or at least impede the widespread adoption of AI, and many of these misconceptions stem from the fact people do not know much about Ai and how it works, this causes a large amount of scepticism around what AI is and just what it can do. A report created by Fintech News (N/A, 2019) showed that users do not trust a decision that is made by an AI and would have a higher level of trust if they were talking to a person, this is because a human will also think about a decision ethically rather than just facts and figures and this is one of the key reasons why people do not trust AI. Why should they trust something that cannot feel and show emotion but just makes decisions based on hard facts?  Once we overcome the issues regarding the actual functional bias of the AI it then comes down to, does the user trust the company which the AI is “technically” working for? Are they ethical? What are they doing with the data they get from all this? This issue can be seen with large companies such as Amazon, betraying the trust of their users (Simmonds, 2020) no wonder users have a hard time trusting Alexa for example. All this is important as machine learning and AI become more prevalent how can we prevent this from causing even more bias in what people think about AI?

How can all this be avoided?

People like transparency and understanding, when people are in the dark they feel threatened and develop distrust understandably, this is why Explainable AI will go a long way to allow people to explain just how the AI is “thinking” and making decisions this will allow users to have insight into how an AI has made a decision and be taken out of the “dark”, this will allow the user to then understand and place more trust in that decision or rationale, as this will be more like just how one human does with another human, this level of understanding will help solidify trust in the system and mean that there is less misconception about the end goal of the AI hopefully alleviating the idea that AI may turn bad. The second way that this bias can all be avoided is by making sure that the companies managing the AI or the ones that are responsible for it follow regulations such as GDPR, data protection act etc. these laws, although not limited to AI will allow the user to know that all the data that is used about them will not be misused and will be held to the highest standards, this will help give the user peace of mind and ensure that the bias they may have when it comes to the companies is lowered allowing AI to move forward.

Where do I stand?

Many aspects of AI have been given a bad name due to the fact people believe that even if AI has good intentions as created by the user, it will still develop non-beneficial ways to get to those goals, however, I feel that in this modern society to move forward, we need to allow AI to help progress us carefully but without bias and because of this view I do feel that I agree with the statement that bias will stop the widespread adoption of AI for a few reasons. Many websites tarnish the reputation of AI and use it as an “enemy” as it provides better publicity and scaremongering meaning they get more money when I feel that the benefits of AI if used correctly will outweigh the negatives because of this bad press and bias however its giving AI a hard time and making it difficult to be accepted into society and therefore stopping widespread adoption. I also feel that a large amount of bias is placed with ill-conceived ideas people hold themselves with no factual backing and the only way that this can be changed is to show and prove with facts and real-life examples just how AI can be used for good in the real world, for example in the medical environment.

Conclusion

In conclusion, I think there is a large amount of bias that currently stands in the way of AI progressing substantially into the modern world, and I think that this is not something that can be fixed easily or quickly as people will need to develop trust within AI and the companies that use it one step at a time, however in time I do believe and hope that this bias can be lessened to allow AI to move forward and enrich our every day lives.