Does AI spell the end of the open internet?

23 Jul 2023

The sentiment on generative AI in the media today veers from awe to apocalyptic. Much of the conversation about either view is based on the potential for big changes in a few years. But the impact of AI is already being felt.

Large language models like ChatGPT rely on massive amounts of data to produce their output, and that data is spread throughout an existing ecosystem of social media networks such as Reddit. Among other platforms, Reddit has recently “wised up” and started to charge for access. This decision is one example of how innovation in one part of the tech world can force changes in another.

These changes can have further knock-on effects: Reddit’s decision to increase the cost of using APIs may result in the exclusion of third-party innovators. The founding culture of the internet has been open source, open access and contributor driven. Rising costs and privatisation of data raise the walls of the internet higher, making it harder for new entrants.

But it also leaves the contributors to these platforms – often users like you and me – cut out of “the action”.

Key questions around data vending (for generative AI or other purposes) remain unanswered: who stands to lose the most – ambitious innovators, social networks or the AI platforms?

Above all, should we who contribute to the platform economy be appropriately compensated?

Much of the stuff we consume online comes from all of us providing our own user-generated content, whether on Facebook, Reddit, Twitter, Wikipedia or many others. Our voluntary generation of all this data provides platforms with a treasure trove for third-party users, which they access via an application programming interface (API).

Recognising the value of this, Reddit has announced its plan to withdraw its public APIs and started demanding payment, despite the protests of its loyal users.

Twitter, a similar but more bite-sized conversation platform, which has been a key input into AI data training models, also started to charge for API access. This new monetisation of previously free or affordable access will influence the overall online ecosystem. While large firms and investor-backed apps can no doubt afford these new charges, most small players cannot. The impact on innovation, fair access, competition and user experience could be vast.

It’s not the first time that Reddit’s content has been repurposed for commercial gain or as content-fodder for the media.

Moreover, there is a question of fair compensation and attribution. None of us expected payment for our photos, jokes or comments, nor necessarily for the labour-intensive and meticulous research of Wikipedians, or free-use photography by the photographers of Pexels that is so widely used, or for that matter any of the artworks, analyses, tutorials and recipes that we put online. Yet we did it out of altruism and perhaps sometimes for recognition, a fundamentally human intent to help our communities and be seen by them.

Read more in Daily Maverick: Real danger of ChatGPT lies in its robbing us of our ability to read and research critically

One study estimated that Reddit moderators carry out more than $3.4-million in unpaid labour each year and send valuable traffic to Reddit through third-party apps like Reddit is Fun and Apollo, relying heavily on open-access APIs that until now were more affordable. Redditors themselves, the real value of the platform, also perform “unpaid labour” when they come to Reddit to engage in a conversation.

As a long-time Redditor myself, it’s not the first time that Reddit’s content has been repurposed for commercial gain or as content-fodder for the media. Content has long been lifted from Redditors and repurposed into strange compilation videos on YouTube that are monetised and generate revenue for their anonymous creators.

Newspapers have found it easy to fill space by recounting opinions and noteworthy remarks from Redditors on particular topics. Exploitation in one form or another has been occurring.

But the response to Reddit’s API move tells us something about a new strength of feeling. It was said that during Reddit’s blackout more than 50% of its subreddits were down, and user engagement dropped. While some users continued to use the site, they “spammed” subreddits with pictures of John Oliver.

I mentioned in a previous article that the productivity paradox occurs when people adjust to the use-cases of new technologies, which sometimes increases expenditure, and delays the impact of productivity gains as new adjustments are made. Sometimes these new technologies face “lock-in”, or they interact poorly with existing technologies that are well established, or the difficulties inherent in the business at scale become apparent further down the line.

‘Dirty work’

For AI, the question really is, while it seems free for now, at whose expense? There is already tiered access for AI – is this necessarily fair?

AI has raised an existential question around compensation on multiple fronts.

Content moderation to flag disturbing content still relies on human labour and human discretion. OpenAI used outsourced Kenyan labourers earning less than $2 per hour to flag disturbing content including sexual abuse, bestiality, murder, suicide, torture, self-harm and incest. While some counselling was offered for workers, they pointed out that they were discouraged from using it due to the high pressure to be productive.

The work’s traumatic nature eventually led the outsourcing firm, Sama, to cancel all its work for OpenAI in February 2022, eight months earlier than planned. There is the risk that the “dirty work” of technology gets outsourced to developing countries, without appropriate compensation and without the social and health support systems to protect and support them.

Read more in Daily Maverick: Artificial Intelligence is moving way faster than the speed of evolution. Could we be in trouble?

While the crypto fallout in 2022 on multiple fronts, including FTX, has cooled affections for blockchain, Web3, using decentralised blockchain technology to ensure that user’s content is permanently associated with themselves rather than housed in a particular platform, remains an interesting potential solution to some of the content and compensation concerns for social media and AI. Web3 evangelists argue that it is possible to compensate users fairly through a decentralised structure, and with strong privacy protections.

But these competing proposals are not yet tangible, and so regulation often fills in the gaps. The state is increasingly drawn into the conversation around AI. Already countries around the world are approaching ChatGPT and its peers with a mix of trepidation and excitement.

Countries such as Italy and China banned ChatGPT in March, but have since lifted such bans. The EU has released the AI Act which aims to ensure safe use of AI and oversight by people rather than automation to prevent harmful outcomes. The UK, ever pro-innovation, is considering “a strategic state” to develop new capacity to meet the challenges of AI, with proposals for an AI regulator to both regulate the market to promote innovation and fair, open access.

We are faced with a powerful new technology, as we have done many times in the past. Nuclear technology can be used for both bombs and for clean power. Social media can be used to manipulate elections, but it can also connect communities in powerful ways. The first innovation of all, fire, gifted to us from Prometheus himself, can burn down forests but it can also cook food and keep us warm. It takes society at all levels to ensure that adoption and use of technology is fair and safe and serves us all. DM

Emma Ruiters is a technology and public policy analyst at the Tony Blair Institute for Global Change.

Categories:

Tags:

DM is much better if you are signed in...

World

‘Dirty work’