Don’t get me wrong, I am not an AI sceptic at all. In fact, I am the complete opposite. I love learning about the innovative ways AI is positively impacting our personal lives, industries and professions and what is in the pipeline to further innovate AI. I was at a virtual event last week where the topic of discussion was ‘Most Outrageous AI Predictions’. One of the predictions was that AI will be built in to future Children’s toys for the purpose of understanding how Children interact with certain toys and how different toys can impact their development. I thought that was quite cool. I also heard another prediction, where companies like Netflix and Spotify can use generative AI to personalise music and TV shows specifically for each individual. The TV shows for instance would have all of your favourite actors, and the plots, settings and outcomes that you mostly enjoy. That would also be very cool, however I wouldn’t class these too predictions as “outrageous.”
However, when I think about ‘uncool’ ways of AI being used, one of my biggest fears is the acceleration of disinformation. We already know the internet is a treasure trove of opinion over fact at the best of times (take Coronavirus and Brexit as a couple of examples in recent years). My general feeling is that with the continued development of generative AI and it’s ease of access by all, the spread of disinformation is only going to get worse.
Once reliable sources may also succumb to the ‘information trolls’ with the help of AI cloned websites, AI-generated articles, AI-generated deepfake videos etc. The list can go on. We have ‘Fact Checkers’ that are supposed to raise awareness of ‘fake news/information’ but I also feel those fact checking sites can also be manipulated with the power of AI.
What is already being done in the fight against disinformation?
The EU as of 2021 started to consider the political, social and governmental implications of AI and was the first international governmental organisation in the world to draft what will be known as The AI Act. As of 9th December 2023, the wording of the act was provisional approved and is due to be implemented in to EU law within 2024.
The EU wants to regulate AI to ensure that innovative development of the technology creates more benefits over risks for EU citizens, countries, and industries. The regulation itself follows a risk-based approach based on how the technology is regulated. Therefore, higher the risk, higher the regulation.
So, what does it mean in terms of disinformation?
The way the law has been structured and worded, generative AI tools such as Chat-GPT would have to comply with transparency requirements. This means any content created through tools such as Chat-GPT, must be published with a caveat stating that the content was created using AI. It also prevents generative AI from producing any illegal content, or publishing summaries of copyrighted data.
The AI Act creates a new governance structure for AI, with one being named as the European AI Board. It would make boards (like this) responsible for ensuring the regulations are strictly adhered too, and that transparency requirements in ‘High-risk’ and generative AI categories are complied with.
However, there is still the issue of disinformation being spread without the use of generative AI. There are well-known practices of groups, hackers, and individuals who purposefully flood the internet with incorrect facts and information to manipulate general populations. When combined with unethical data tracking, you can turn the minds of the general populous with just a few sentences. This was seen during the UK’s Brexit in 2016 and the US General Election in 2016. This is not a conspiracy, this is proven fact. Just look at the story in Time magazine of how Cambridge Analytica and Facebook could well have been used to manipulate the American public in the 2016 federal elections (here).
How can we really trust facts and information, and use AI as defence in the fight against disinformation?
With a mix of interest and fear in this subject, I turned to my network of experienced professionals and experts to ask what they thought the solutions were. I was mostly interested in how we can trust the fact checkers and what we could use to minimise the risk of wide-ranging disinformation. I got a few responses which amounted to ‘sucking in air through teeth’ and words in the realm of ‘it seems impossible’. However, in a couple of other conversations I was discussing the possible use of Blockchain and AI combined as a feasible defence.
An article published by the Harvard Business Review in 2021, went in depth into this very subject of using Blockchain. The subtitle contained the words “It’s not a cure-all, but it does have the potential to address many of the risks and root causes.” I was intrigued as I had only known Blockchain’s application in the financial industry, but it made sense in this context also.
“Blockchain systems use a decentralized, immutable ledger to record information in a way that’s constantly verified and re-verified by every party that uses it, making it nearly impossible to alter information after it’s been created.” Kathryn Harrison and Amelia Leopold (2021, HBR)
One of the biggest challenges that we face (even with the introduction of the EU’s AI Act), is that it is hard to combat disinformation as there are “no standards for identifying, labelling, tracking, and responding to manipulated media across digital platforms.”
Blockchain seems to be a valid technological opportunity in finding a solution to a huge problem we have in the world of information. With the addition of AI, we could much easily identify deep fakes, incorrect facts, and false information and use a system which records everything so we can combat and target those who are spreading disinformation. I believe we have a huge fight on our hands which will only get worse. With the US and UK governmental general elections looming in the near future, how are we going to trust that what we are reading about competing parties is true?
If anyone reads this and wants to discuss this further, get in contact with me or NUPROJEX.