You joined social media to communicate — when did you agree to let an algorithm choose what you see?

There was once a fairly simple bargain at the heart of social media.

You joined a network.

You chose people to follow.

They chose whether to follow you.

You published photographs, opinions, news, invitations or observations.

Other people did the same.

It was, fundamentally, a communications system.

Somewhere along the way, something changed.

The social network ceased merely carrying the information we had chosen to receive and increasingly began deciding what information we should receive.

The distinction is profound.

And Australia is now proposing that users should be given some of that decision back.

Under the Federal Government's proposed “My Feed, My Way” reforms, social-media platforms would have to give Australians over 16 a genuine choice over their default feed.

A user could choose personalised material recommended by an algorithm.

Or the user could choose to see material from the friends and creators they have actually decided to follow.

That sounds almost revolutionary.

Perhaps it shouldn't.

Perhaps it is simply returning social media to one of the principles upon which people thought they were using it in the first place.

We joined because we wanted to communicate

Think about why people originally opened social-media accounts.

A person joined Facebook to communicate with friends and family.

Someone joined Instagram to share and view photographs.

A professional joined LinkedIn to maintain business relationships.

A person joined X to follow politicians, journalists, organisations and people they found interesting.

The transaction seemed reasonably clear.

Join the network and communicate through it.

But modern social media increasingly involves another transaction.

The user supplies attention.

The platform decides where much of that attention goes.

What appears on the screen is therefore no longer necessarily the information the user explicitly requested.

It can be information selected because a computer system predicts that the person will watch it, click it, comment on it, argue about it or remain on the platform because of it.

That is an entirely different proposition.

There is nothing inherently wrong with recommendations

Algorithms are not intrinsically sinister.

They solve a genuine problem.

There is vastly more information available online than any person could possibly consume.

Recommendation systems can discover music we enjoy, introduce us to writers we have never encountered, find businesses we might need and present news about subjects that interest us.

Sometimes the algorithm is remarkably good.

People may actively prefer it.

That is precisely why Australia's proposed reform is important.

It does not require people to abandon algorithmic recommendations.

It proposes giving them a choice.

If someone wants a platform to search billions of pieces of content and decide what might interest them, that should remain available.

But another person may say:

I already made my choice. I followed the people and organisations whose information I want to receive. Show me them.

Why should that be an unreasonable request?

The business model complicates the relationship

Social-media platforms are businesses.

They employ people, operate enormous computing systems, maintain global infrastructure and provide extraordinarily sophisticated services.

Someone has to pay for that.

For many platforms, advertising has provided much of the answer.

Attention consequently has commercial value.

The longer people remain on a platform, the more opportunities there are to display advertising and generate revenue.

This creates an obvious tension.

What the user wants and what maximises engagement are not necessarily the same thing.

A user may want to check photographs from family members, read updates from three organisations and leave.

A platform may have a commercial incentive to keep that person scrolling.

Recommendation systems help bridge that gap.

There is nothing surprising about a business attempting to increase use of its product.

But users should understand the transaction.

The apparent price of a free social-media account may increasingly include something more valuable than money:

a portion of control over our attention.

People aren't necessarily asking for less information

This distinction is frequently lost.

People who object to algorithmic feeds are not necessarily asking to be protected from information.

They may not want government deciding what they can read.

They may not want a technology company deciding either.

They may simply want to make that decision themselves.

That is not censorship.

It is selection.

Every person already selects information constantly.

We choose which newspaper to read.

We choose which television program to watch.

We choose which podcast to hear.

We choose which book to buy.

We choose which conversation to join.

And traditionally, on social media, we chose whom to follow.

The principle should not suddenly become controversial because a computer can make the choice for us.

The right to speak is not the right to occupy someone else's attention

There is a broader principle here.

Freedom of expression matters enormously.

People should be able to express unpopular opinions, challenge governments, disagree with prevailing views and participate in public debate.

But freedom of expression does not mean every person is entitled to an audience.

The right to speak does not create a right to occupy someone else's attention.

That distinction becomes increasingly important in an algorithmic world.

If someone publishes an opinion online, that is expression.

If I choose to follow that person and read it, that is my choice.

If a platform repeatedly inserts that person's material into my feed despite my never having chosen to follow them, another decision has occurred.

The platform has exercised editorial power over my attention.

It may have perfectly legitimate reasons for doing so.

But it is still exercising power.

Algorithms are editors, even if we don't call them editors

A traditional newspaper editor chooses which stories appear on the front page.

A television producer chooses the running order of a bulletin.

Those decisions are readily understood as editorial decisions.

Algorithms perform something conceptually similar at extraordinary scale.

They rank.

They suppress.

They recommend.

They repeat.

They decide what appears first and what may never appear at all.

The important difference is that the algorithm can make a different editorial decision for every individual.

Two people can open the same social-media platform at exactly the same moment and effectively enter different information environments.

Australia's eSafety Commissioner has acknowledged that recommender systems can produce both positive and negative outcomes and that engagement-based systems need not be the only model.

The debate therefore should not be reduced to whether algorithms are good or bad.

The better question is:

Who ultimately gets to choose?

There is an important difference between discovery and delivery

Social networks could easily preserve discovery.

There can be a recommendations page.

There can be a “For You” section.

There can be trending subjects.

There can be suggested creators.

There can be prompts asking whether we would like to discover something new.

Those functions can be useful.

But discovery does not have to replace the feed a user deliberately constructed.

A person could have both.

One feed could effectively say:

Here is what you asked for.

Another could say:

Here is what we think you might like.

The distinction would be transparent.

More importantly, the choice would belong to the user.

Australia is proposing exactly that choice

The Albanese Government's proposed Digital Duty of Care legislation now puts this principle squarely into Australian public policy.

Social-media platforms would have to notify new and existing users and offer a choice over their default feed.

Users who want personalised recommendations could opt in.

Those who do not could instead see the friends and creators they have chosen to follow.

Prime Minister Anthony Albanese described the principle succinctly when announcing the proposal:

“It's not about giving government control. It's about giving people control.”

That is the strongest argument for the reform.

It does not require government to determine which political opinions Australians should encounter.

It does not require government to construct our news feeds.

It does not prohibit algorithms.

It gives the decision to the individual.

There will, of course, be legitimate arguments about the details of the legislation, how platforms implement it and how broadly the government's proposed digital duty of care should extend.

Those debates should occur.

But the principle of meaningful feed choice is difficult to dismiss.

Choice must also be genuine

There is one qualification.

Providing an option somewhere deep inside a settings menu is not necessarily meaningful choice.

Nor is giving users a choice and then repeatedly steering them back towards the platform's preferred setting.

The government's proposal recognises this by referring to a genuine and enduring choice.

That matters.

A user who selects a following-only feed should not have to rediscover the setting every week.

The platform should remember the decision.

Equally, users who later decide they miss recommendations should be able to turn them back on.

Real choice works both ways.

Social media would still be social media

None of this requires dismantling social networks.

They would remain enormously powerful communications and discovery systems.

Algorithms would remain available.

Advertising would remain.

Creators could still find audiences.

People could still discover strangers with interesting things to say.

The change is much smaller than that.

It simply establishes a boundary.

The platform can recommend.

The individual can decide whether to listen.

The Times View

The great achievement of social media was that it gave ordinary people extraordinary communications capability.

Anyone could publish.

Anyone could connect.

Anyone could follow voices previously inaccessible to them.

That achievement should not be understated.

But somewhere during the evolution from social network to attention economy, the relationship changed.

We chose whom to follow.

Then the platforms increasingly chose what we would see.

Those are not the same thing.

Algorithms can be enormously useful. They can introduce us to ideas, people and information we would otherwise never discover.

People who value that service should be free to choose it.

But people who would rather construct their own information environment should be equally free to decline it.

Australians do not necessarily want to be presented with less information.

They want greater authority over the source of the information presented to them.

And beneath the technology lies a principle much older than social media.

Everyone should be free to speak.

Everyone should be free to publish.

Everyone should be free to persuade us to listen.

But the right to speak does not create a right to occupy someone else's attention.

In the digital age, our attention has become an extraordinarily valuable commodity.

Perhaps ownership of it should remain with us.