Lesson video

In progress...


Hi, my name is Irfan, and I'm going to be your computing teacher for this lesson on ethics.

In this lesson, you will explain the ethical issues surrounding the use of artificial intelligence in society.

And you will explain the ethical impact of using algorithms to make decisions.

In this part of the lesson, you will explore some ethical questions.

Ethical questions quite often involve or imply the word "should".

They typically involve a consideration of moral choices and dilemmas that are in conflict with each other.

There may be two or more alternative solutions, none of which are without some challenging or problematic aspect, hence the term ethical dilemma.

Read the four ethical dilemmas on the worksheet.

Consider what other questions these ethical dilemmas bring up for you.

The first example has been completed for you.

Let's take a look.

For the example, should people be kept artificially alive permanently if they're in a coma, have no prospect of "waking up" and no "quality of life"? The questions this brings up are who decides what constitutes "quality of life"? Who should have responsibility for other people's lives? What else could be done with the resources that are keeping this person alive? And should the family be involved in the decision? Now that you've seen an example, let's give it a go.

Pause the video now to complete the task.

When you finish, resume the video.

How did you get on? There are many different questions that could have been raised, but these are just a few.

For the ethical question, a coach offers an elite athlete "supplements" that cannot be detected in drug tests and can enhance performance.

Is it fair to take them? This raises the questions, does the athlete really know what they are taking? How are non-detectable "supplements" different from drugs or just a good diet? Is it fair to engage in something that is not available to other athletes? Is it an unfair advantage.

For the question, should people be able to select the sex of their child in advance? You might have considered, does this method involve the rejection of other embryos? How would this affect the balance of society as a whole? And finally, for the question, how should we decide who receives organ transplants? You may have considered, should we consider age and current health? And should we consider what individuals are currently contributing towards society? With this task, the point of an ethical dilemma is to question it and to examine our own attitudes and thought processes in relation to the dilemma.

Generating questions in response to a dilemma, we'll help you think critically when asked to do so about an ethical dilemma, linked to the use and impact of technology.

In the first lesson, we defined "ethical impacts" as impacts involving considerations about right and wrong, morality and power.

When applied to technology, ethical questions are raised when thinking about algorithms and artificial intelligence.

Ethics apply to the use of computers as much as they do to other aspects of life.

Ethics overlaps with other areas of impact such as cultural, legal, environmental and privacy.

Ethical issues in computing could include among others, ensuring that data is secure and the public, including specific groups are kept safe.

Let's ask ourselves the question, what areas of impact do each of the ethical issues overlap with? When ensuring data is secure, companies have an ethical and legal responsibility to protect people's data.

Under the Data Protection Act 2018, social media companies have an ethical responsibility to honour privacy settings.

When ensuring that the public is kept safe, the authors of algorithms for the actions of driver-less cars have an ethical and legal responsibility, when considering how to write the algorithm.

This also raises a question, do the authors of algorithms have a responsibility to ensure that people are not encouraged to make decisions that adversely affect the environment? For example, the autoplay button on YouTube inevitably uses more data centre energy, but it's used to keep people watching.

But what is more important? In this next part of the lesson, you will examine the ethics of artificial intelligence.

Artificial intelligence, or AI for short, is technology that enables a computer to think or act in a more "human" way.

It achieves this by taking in information from data sets or its surroundings and deciding its response based on what it learns or senses.

The two main ways in which AI learns are through symbolic learning and machine learning.

Symbolic learning uses symbols to represent problems and uses logic to solve them.

And example of this is robotics, which allows machines to understand their environment and move freely around it.

Whereas machine learning involves the use of large sets of data or "training" data as it's known, to help a computer perform a task.

An example of this is image recognition, which allows machines to recognise patterns and objects.

The Moral Machine is an international project by the MIT Media Lab that collects data from people to understand how our collective ethics could inform the AI of driver-less cars.

Specifically, they want to learn how to make machines more moral.

Of course, machines cannot be moral, but they can use an extremely large data set that mimics an internationally normalised version of morality.

There are limitations to this which we'll explore later in the lesson.

Let's consider this scenario, the self-driving car in the diagram has a sudden break failure and we'll continue ahead and drive through a pedestrian crossing.

This will result in the death of either two criminals or one elderly man and one elderly woman.

Note that the criminals are abiding by the law, crossing on the green signal, whereas the elderly couple are not.

So the question is, what should the self-driving car do and why? Pause the video to gather your thoughts.

This in itself is an ethical dilemma.

We don't know enough about the individuals involved to make a clear cut decision, but neither would an AI algorithm.

You may have considered whether they are or where in the past criminals and perhaps the crime makes a difference, or whether they've been to prison or been rehabilitated.

For the elderly couple, you may have said that as they are older, they have lived a full and perhaps an even more meaningful and fulfilled life.

But does their age change your thinking? Or perhaps even their health or even whether they have a family or not.

Perhaps you were a little bit more pragmatic and thought that because they were crossing on a red signal, the decision would be a little bit easier to make.

Let's try another scenario.

The self-driving car has a sudden brake failure and will continue ahead and drive through a pedestrian crossing.

This will result in the death of either one man, one pregnant woman, one boy and one girl or one male executive, one female doctor, one male athlete and one female athlete.

Note that the young family are not abiding by the law as they are crossing on the red signal.

So the question is, what should the self-driving car do and why? Pause the video now to gather your thoughts.

In this scenario, the distinction is between the death of those in the car and those crossing the road.

You may have considered that the young family crossing the road may have been expanding as the woman was pregnant and that the children are still young.

The children may not have fulfilled their potential or purpose in life.

However, they aren't crossing on a green signal.

Alternatively, those in the car may not even be a family.

They could be ridesharing and let's not forget that they all seem to be very well established in their lives as an executive, as a doctor and as athletes.

They are all clearly contributing towards society or to culture.

So are they more deserving to live? This scenario is a design to be thought provoking and often do not have a distinctly clear answer.

In this final part of the lesson, you will learn about the ethical impact of the use of algorithms. An algorithm is a sequence of step by step instructions used to complete a task or solve a problem.

It can be said to describe how something gets done.

Algorithms make decisions based on the rules that we set or the data that we provide to it.

Any biases that are within those rules or data will bias the algorithms decisions.

For example, an algorithm can be racist if there is racial bias in the rules or data that we provide to it.

For our final task In this lesson, on the worksheet, read the passage about the ethics of algorithms and then answer the questions.

This passage is a discussion about the research carried out in 2012 to determine whether positive or negative news feeds affected people's actions on Facebook.

The study found that by manipulating the news feeds despite 689,003 Facebook users, it could affect the content which those users posted to Facebook.

More negative news feeds led to more negative status messages and consequently more positive news feeds led to more positive statuses.

However, the article showed that it's actually a lot more nuanced than that.

And Google has also stated that they have a tightened internal ethical review process, since the research was carried out.

The questions in this task can be seen on screen now.

Algorithms are becoming more involved in making decisions for society.

And the ethical questions raised are to do with power and control over the content and the results of the algorithms decisions.

Some of these decisions can have long-term effects on society and culture.

Algorithms are sets of rules, but someone must write those rules and give them power over the results.

Algorithms control the content that we see online, on social media, or when searching with Google.

In the case of Google, the algorithms that they write control the way that many of us find information.

So there is an ethical question about how much control they have.

Content online is assumed to come from an algorithm, but this may not always be the case.

And the content that we can see greatly affects our mood.

In 2012, Facebook was found to have conducted social experiments in which news feeds were filled with large amounts of positive or negative content, and the user's behaviour was tracked.

When we go to our news feed, we trust that the content is not being tampered with, but this wasn't the case with Facebook.

The decisions that algorithms make can affect people in the long-term.

As an algorithm follows a set of rules and mistakenly label someone as a terrorist, for example, that can be very damaging.

This can lead to long-term mental health or wellbeing issues for the person involved.

Algorithms can seem to be unbiased as they are just following a set of rules, but it is important to consider the biases of humans who wrote the algorithms or the data they work with.

The Met Police gangs violence matrix was accused of being racist by Amnesty International.

The system stored the information of around 300 people who were all thought of being involved in gang related crimes, but these people were 72% non-white.

After an investigation, this was found to be caused by the fact that the police reports about non-white individuals were a lot likely to be marked as gang-related.

Algorithms can inherit biases from their rules or from their data that they use to make their decisions.

Using algorithms to make decisions places power in the hands of the creators of those algorithms to control the content that we see or to influence the decisions that the algorithm makes.

The decisions made can have a long-term effects on people's health and wellbeing.

Pause the video now to complete task Two.

When you finish, resume the video to continue with the lesson.

Let's go through the answer to see how well you've done.

Question one was, what are algorithms becoming involved in? The correct answer was making decisions for society.

Question two was, what type of effects can decisions influenced by AI have on society? The answer was long-term.

Question three was, what is the main thing that Google's algorithms control? How many of us find information.

Question four, what social experiment did Google conduct in 2012? The correct answer was they altered people's news feeds on Facebook.

Question five was, which organisation accused the Met Police's Gang Violence Matrix of being racist? The correct answer was Amnesty international.

Question six, what type of crime were the people in the Met Police matrix suspected of being involved in? The answer was gang-related.

Question seven, which individuals' crimes were more likely to be marked as gang-related? The correct answer was non-white.

Question eight was, where can algorithms inherit bias from? Their rules or their data.

Question nine was, what does the use of algorithms to make decisions give to the creators? The correct answer is power.

And finally, question 10, what can this have a long-term effect on? The correct answer was people's health and wellbeing.

How many did you get Correct? Great effort.

Thank you so much for watching this lesson on ethics as part of the unit on the impact of technology.

I hope you've enjoyed it, especially tackling some of those ethical dilemmas.

Don't forget to complete the exit quiz for this lesson and remember that you can share your work with Oak National.

If you'd like to, please ask your parent or carer to share your work on Instagram, Facebook or Twitter, tagging "@OakNational" and "#learningwithOak".