Fake news and fake information have been big issues for years now. Both have had an impact on domestic politics in several countries, but also been linked to international politics, especially through Russia’s endeavours to cause political problems in the West. Some social media platforms have tried with marginal success to filter content on their services. But generative AI can take the fake information arms race to the next level.
Fake information vs real information
Trolls and troll factories have been an important source of fake news. Of course, some political publications and ordinary internet users have played a significant role in modifying real information and creating and distributing fake information. It is not so easy to differentiate between real and fake news, and it is often difficult to detect original sources of fake news.
AI used to detect and filter fake news
AI has been used to detect and filter fake news. But now we are seeing the other side – AI can also generate content that is not true.
TechCrunch told how some companies have started to use generative AI to create content and web sites. This is a new way to generate spam and get greater visibility in search engines. Basically, it enables you to rapidly create loads of content about a topic. It also makes it possible to link to others and achieve good page rankings.
This is not necessarily illegal, but it can make it much more difficult for users to find real content and real products. In a way, it challenges how we have been able to find information, services and products on the Internet. Search engine companies are working on solutions to combat this, but it is not easy when the area is now developing rapidly.
A new level of fake information
Basically, the same can happen with fake information and news. There have already been worries about how AI development enables the creation of videos and speeches that are not authentic and can include fake information, commonly known as deepfakes.
Fighting against individual websites and videos with fake information is one thing, but when AI generates hundreds, thousands or millions of sites that include the same fake information that also links to other fake information sources, it takes the fight to a new level.
Basically, we can say that quality of information is more important than quantity. Imagine we have one hundred websites that contain true information on something – then we have a million sites that have fake or twisted versions of the truth. This can influence many people.
How to detect real things
In a way, this creates a new situation on how to find reliable information online. One answer can be to trust well-known and respected media and information brands. But even there we have seen conspiracy theories espoused. This is how these “most respected media brands,” run by the world’s media elite, manipulate us. So, even established and once-respected media outlets can’t be trusted to deliver the truth.
We need to find technological solutions for this. We can compare this to fraud detection solutions (e.g. credit card and telecom companies use them). They typically have two different approaches to detect fraud transactions: 1) identify atypical transactions, and 2) identify known fraud methods.
With fake information, this would mean having solutions to identify information and a source that:
1) Includes something atypical, for example –
- Information that is not in line with most other information sources;
- Websites that have been created most recently;
- Information sources that link to each other.
2) Includes information or structures that are typical for fake information –
- Those that include typical propaganda, political influence or conspiracy patterns;
- That comes from certain unreliable sources;
- Websites and link structure that resemble new, artificial information sources.
Those are just some examples of how we could start to fight against this kind of systematic way to create fake information. Probably it is not enough, but it already illustrates that we can soon have totally new challenges to find information on the internet.
Who is involved?
As the TechCrunch article illustrates some companies already use AI to generate content that is for their marketing and visibility purposes. Some companies can also be interested in creating fake information, for example, fake test results and reviews of products.
At the same time, fake information is often linked to political purposes. Sometimes this means domestic politics, but we have also seen how important it has become in international politics, even in hybrid wars.
Internet companies, like Google, Meta and media companies (no one knows what Elon Musk and Twitter want to do), need to invest in new solutions to give more reliable information and search results to users. But it is quite clear that this area is so critical that governments would leave it only to private companies.
Role of governments
Governments most probably want to create technology and also regulation for this area. It is also important to educate people to understand what to trust and not trust. And in many countries intelligence and security services have become active in this area too. People can benefit from managing fake information themselves. They can use their own brains, check facts and not distribute unreliable or fake information. But we already know it is not enough. So, this will become a new battlefield to make AI to generate fake information, and at the same time to detect and remove it. And it will also be a part of the new cold war between countries and blocks.
Related article: AI may balance the fake news arms race, but it can’t cure gullibility
Be the first to comment