We have mentioned how the ethics of AI may affect the future of AI. In this lesson we will explore this in more detail. 

For the companies who are creating AI programs, ethics issues are very important. In 2023 the level of abilities of generative AI has rapidly increased, and this has led to many experts in the field to advise that we limit the progress of research to make sure that we fully understand the implications of AI. This would hopefully mean that we don’t create unfair or dangerous AI programs.

Governments around the world are beginning to create laws and legislation that say what AI companies can and cannot do.

​

It is also important for the users of AI to be aware of unfair and dangerous outcomes for using AI. Here are some issues that we can be aware of.

• Bias and Fairness: AI systems can inherit biases from the data they are trained on, which can lead to unfair outcomes. It's important to consider how AI systems can be designed to be fair and unbiased, and that all systems should be tested before use.

• Privacy: AI systems often rely on large amounts of data, which can include personal information. This data must be stored securely and not transferred to other companies or for other uses that weren’t agreed.

• Transparency: Although AI systems are not entirely predictable, it is important for AI companies to explain how their systems make decisions and who is responsible for those decisions. This information should also be passed on to people whose data is being used by someone running the program. 

• Accountability: There should be easy and effective ways for users of AI to report misconduct or errors in the system. And for anyone who’s data is being used by AI to be able to remove their data from the system. If real harm has happened, anyone should be able to take legal proceedings against an AI company.

• Safety and Security: AI systems can have real-world impacts, so it's important to make sure they are designed to be safe and secure. This could include consulting and testing the systems with experts and the public to uncover any impacts that the design team have not considered.

• Informed Consent: It's important for people to understand how AI systems are being used and to have the opportunity to consent to their use. This is particularly important in sensitive areas like healthcare.

• Social Impact: How will AI be used to benefit society as a whole.

 

DISCUSSION

This is a short list of considerations that are important in the ethics of AI. It's a complex and evolving field. Can you think of any others? In pairs, choose one of the points above and think of an example where you would not want an AI program to be used with your data. 

​

For example with Privacy, you may not want an airline company to use your flight data to recommend holiday destinations for you, as that may feel too personal. Others could find this a useful feature, so discuss with your partner and see if you agree on these points.