Nowaday, artificial intelligence (AI) enjoys immense power, playing a big role, sometimes secretly, sometimes openly, in how people and society do their everyday activities.
Although AI has tremendous power to transform the lives of people, it also has several limitations and ethical issues regarding how technology will be used if it falls into the hands of wrong people or organisations.
How AI Is Used
- Machine learning, automation, robotics, and sensors are some of the powerful forms of AI that are being deployed by companies at an unprecedented scale.
- AI technology promises to be cheaper but efficient. Organisations are using AI to reduce costs and bring efficiency.
- Simple-repetitive tasks that don’t require high-level decision making are no longer being done by humans. AIs are performing these tasks.
- Thanks to more powerful computers and bigger data sets, AI has also become more sophisticated, capable of performing tasks like a highly-skilled individual.
- In health and diagnosis, AI is used to generate bills and do all the paperworks. AI can analyse data so fast that in the near future, it will be easy to make an accurate treatment decision, as all the medical knowledge on a disease will be available with a click of the mouse.
- When it comes to employment, AI technology is being used to process resumes. It also helps to find suitable candidates for jobs by analysing the voice and facial expressions of candidates.
- Today, AI is used by Amazon and Google, and industries like banking, healthcare, retail etc.
- Algorithms power Google and Amazon’s search queries.
- Banks are also using AI to objectively assess creditworthiness of individuals and small businesses, eliminating human biases.
- Organisations are using AI to manage sourcing of materials and products.
- In sectors like pharmaceuticals trial and error of products are being done by AI at a fraction of earlier costs with more efficient end-result.
- In some sectors, AI is not replacing workers, only making their tasks easier and efficient. Workers become more valuable to employers as they become more productive when they use AI. However, the same is not true with sectors where jobs mostly involve data processing.
- However, it is too early to see it will eliminate entire jobs in certain categories. We don’t know yet whether AI will completely replace highway toll-takers.
Arguments In Favour Of AI
- It enables workers to do more work in less time, make lesser errors, and become more efficient.
- Thanks to AI, businesses are gaining new insight into key details like sales, invoices, ordering and other key information related to finance and customer behaviour. They don’t have to hire experts to discover how their businesses are doing, and what are the problem areas that are affecting the bottomline of the company.
- Banks are using AI to get an authentic picture of creditworthiness of an individual and small businesses. Informed assessment could be made in minutes, eliminating the need to do the cumbersome paperworks.
- Supporters of AI’s algorithmic decision-making ability argue that unlike human decisions, it will not be biassed, prejudiced and subjective.
Ethical Concerns Of AI
- Experts have highlighted three major ethical concerns associated with AI. These include: Bias and discrimination, role of human judgement, privacy and surveillance.
- Google is widely considered as a pioneer of using algorithms in their search results and other tasks. But, algorithms have been found to be problematic in certain cases.
- We have heard of many instances of AI software being used to predict future criminals and how they ended up showing biases against black people.
- Without human intervention, AI technology could not tell the truth. Neither can it protect confidentiality.
- AI chatbots and algorithms have not trained to be true. They have been trained to be plausible. We don’t know how much of what the bots generate is true, and how much of it is false or plagiarised.
- What is the source of AI’s knowledge? What kinds of data have been entered to make chatbot behave like an intelligent human being?
- Does an organisation or a professional damage the trust others place in them, when they use AI to perform the tasks assigned to them. We don’t know whether a freelance writer has used AI to create a so-called original copy, or whether it has really been written by a human being.
- Organisations and individuals may start offering financial and other advice to others using AI. But information generated by AI may be inaccurate, false or plagiarised. Simply cutting and pasting what the bot generates, and not doing research on what the chatbot has generated, will cause harm to others who may act upon this false information.
- Algorithms have inherent biases. Who should get a particular job, who should be given bank credit, who should get parole. When it comes to taking such decisions, AI can go against women, vulnerable groups like black people in America and scheduled castes and scheduled tribe people in India. AI can replicate biases that already exist in our society against women and minorities.
- If AI starts relocating human biases, such biases will get scientific credibility. People may say, look, even machines don’t want to hire certain people.
- With virtually no government control and regulation, private companies, powerful big tech organisations are using it to know the medical and records of individuals, behavioural patterns, buying habits without paying any attention to the privacy of individuals.
- Can AI software outthink the human brain? Or, do we still need a certain amount of human judgement, especially when algorithm-based technology perpetuates biases that exist in our society?
- Banks need to be very careful in using algorithms to assess creditworthiness of a loan applicant, as algorithms have a higher chance of favouring certain classes or going against certain classes.
Ethical Issues Of Unemployment
- AI has been steadily replacing entry level jobs being done by humans. Advanced AI is automating tasks previously done by humans.
- We can understand the ethical issue of employment with an example. If self-driven trucks can reduce road accidents, nobody will argue against its use. But what will happen to drivers if they are replaced by technology?
- The same thing could happen to office workers, if their works are replaced by chat bot or other AI technology. True, AI could accomplish certain tasks better than humans, but jobs lost to superior technology could create widespread unemployment.
ChatGPT And Ethics
- Ethical issues related to AI have been brought to the fore with the launch of ChatGPT, an AI chatbot, by Open AI. ChatGPT has divided experts like never before.
- Some hail it as a technology that could revolutionise the internet, others are concerned this technology could be used to spread hatred, or spew venom against vulnerable people.
- It has also been proved that it could also be misogynistic at times.
- ChatGPT could be used by university and school students to plagiarise their homeworks and research studies respectively.
- ChatGPT also has racial and gender biases built into its technology that could be hard to control by individuals. It could throw up worrisome content against a particular gender or race.
- Despite the tall claims made by Open AI, ChatGPT may not be as accurate as conventional search.
- ChatGPT is trained on a huge amount of dataset of text. If that data is not accurate, or is biassed, then it could also throw up unfair, offensive and/or inaccurate results.
- At the moment, we are not very sure how diverse the dataset is. If the dataset is representative, then bias and inaccuracy may not be a big problem.
- Since ChatGPT has been trained to generate human-like text, this technology could be used for spreading fake news.
- Safeguards need to be put in place, so that it is not used by terrorist organisations.
- Many experts have also raised concerns that it could be used to impersonate individuals, thus ruining the reputation and careers of innocent individuals.
- Many experts have termed the models like ChatGPT “stochastic parrots”, because they just spouted words without understanding them. Similar to parrots, ChatGPT does not understand what it says, nor does it care.
Will AI Destroy The Human Race?
- In a recent research paper, Google and Oxford scientists have highlighted that not only AI is likely to wipe out humans but it is also almost inevitable once machines acquire advanced intelligence.
- AI systems can wreak havoc if used maliciously.
- AI-enabled technologies could behave in the way that has not been intended by its creators.
- So far, human beings are known as the most intelligent beings on earth. Physical and cognitive tools available with us humans helped us get the better of stronger and faster wild animals. We had created tools to control them.
- Will AI, one day, become the most intelligent entity in the world, and destroy humanity? How to prevent AI turning into a monster that we can’t control is the biggest concern of AI at the moment.
Way Forward
- Policies must be put in place that ensure Government and private entities use AI to benefit humanity as a whole.
- AI should be human centred. Like any technology, people should benefit from AI, not the other way around. Technology should not be allowed to have untrammelled power over our life.
- AI bots are getting better and becoming sophisticated when it comes to replicating human conversation. Its speed and ability to process massive data are beyond any doubt, but it should not be trusted to be fair and neutral all the time.
- Regulatory frameworks should be in place to ensure that AI systems should not be misused against certain sections of the society.