To listen to some people, you’d think that society was at risk of falling into a “Dark Ages,” driven by AI being used for disinformation. 

Without a doubt, disinformation is an issue that is now running through society, and AI can be enormously effective and creating and spreading that. However, what we shouldn’t lose sight of is the human role in monitoring and managing AI, because that is often where the gap is that’s facilitating the flow of bad information.

When AI Is Not Used For Benign Purposes…

The benefits to society of embracing AI and many and manifold. Just some of the ways that AI is being used to create more informed and capable societies include:

• Advancing Healthcare and Drug Discovery: AI algorithms analyse vast medical datasets, accelerating drug discovery and personalised medicine. Predictive analytics optimise drug manufacturing processes, leading to more efficient treatments and improved patient care. For example, AI is already better at detecting some cancers in an early stage than specialist doctors.
  
  
• Addressing Gender-Based Violence and Harassment: AI tools monitor communications for inappropriate content, creating safer spaces and amplifying voices against abuse. For example, AI is being deployed in gaming spaces to ensure that all players can enjoy their experience in a safe and comfortable environment. 
  
  
• Fighting Human Trafficking: AI-driven image recognition identifies suspect advertisements on traffickers’ websites. This aids law enforcement agencies in their fight against human trafficking, and protecting vulnerable individuals.
  
  
• Enhancing Disaster Response: AI-powered predictive models analyse historical data to predict natural disasters. For example, Japan, which is one of the most prone nations to natural disasters on the planet, is leveraging AI to improve its disaster readiness and response processes. 
  
  
• Improving Weather Forecasting: Meteorologists increasingly rely on AI algorithms to enhance weather forecasting accuracy. Accurate predictions benefit agriculture, transportation, and disaster preparedness.
  
  
• Promoting Sustainable Practices: AI supports just about every level of sustainability, including smart power grids and efficient cities. One area of particular interest, however, is precision agriculture. Here AI helps to minimise disease and pests in plants, ensure the overall health of farms, and ensure good quality yields that support both the farmers and their communities.

These are just some of the ways where AI is helping to drive positive societal change by producing better quality information, analysis and insights, more efficiently and productively than would have been possible through human insight alone. 

And yet there is a dark side to AI as well.

“The rise of AI fake news is creating a ‘misinformation superspreader’” wrote The Washington Post late last year, noting that websites hosting AI-created false articles had increased by more than 1,000 per cent. 

“AI is making it easy for nearly anyone — whether they are part of a spy agency or just a teenager in their basement — to create these outlets, producing content that is at times hard to differentiate from real news,” the article reports. 

This is, without a doubt, a problem. As noted by an AI expert, Wasim Khaled, in a CNET feature, AI could erode trust and our “shared sense of reality.” If we reach the point where we don’t feel like we can trust any information, or allow ourselves to be “educated” by false information, the collective ability of society to share ideas, communicate, and solve problems is undermined. 

That could in turn see us experience a cultural decline. Just as the Catholic Church was allegedly a catalyst for the medieval Dark Ages by dominating the knowledge of communities in line with its political and spiritual ideology, so too may our society’s collective intellect become subservient to propagandists pushing falsehoods. 

This can manifest itself in many ways, from undermining the character of a politician (eg: when an image of an Australian politician was altered to give her a bared midriff and larger breasts), right through to supporting the grandest of conspiracy theories (as noted by Vice, the next Qanon, society-disrupting conspiracy theory could be entirely the work of AI). The result, across all of our communities, will be less well-informed and educated people, greater incidences of extremism breaking down cohesion, and a constant burying of the truth. 

But is AI the responsible party here, and are we willing to trade the potential benefits of AI – including in areas that we haven’t even considered yet – in an effort to clamp down and minimise the use of AI in deception?

The answer is no. AI may well herald in a Dark Ages, but preventing that comes down to how humans shepherd the technology. If we do it well, we will have a second enlightenment instead.

Shepherding AI to positive information flows

Most of us want to use AI responsibly and in service of human progress. In doing so, we need to come together collectively to be stewards of good information and actively combat misinformation and deep fakes. 

Some of the steps that we can all take – as both organisations and individuals – are:

1. Educate and empower users. Helping people understand how to combat AI-driven disinformation will help undercut the reach and power of that disinformation. This means developing and providing tools and resources to help users verify the source, authenticity, and accuracy of online content.
   
   
2. Develop and deploy detection and moderation systems. AI is also a powerful tool for fighting against bad AI. We can build tools that use advances in computer vision, natural language processing, and machine learning to detect and flag AI-generated content. Once flagged, human moderators can take steps to verify where that information comes from, and better characterise when the information contains falsehoods.
   
   
3. Adopt and enforce ethical and legal standards. On a very simple level, while AI is of invaluable assistance in generating content, audiences will continue to have trusted sources. Those sources have a moral and ethical obligation to adopt and enforce ethical and legal standards for the use of AI in content creation and dissemination. This means being transparent when AI is used, and additionally ensuring that human oversight remains over the content produced, to guarantee that it is quality, accurate information.
   
   
4. Take responsibility for the content being published. In an era where social media allows everyone to publish material, those that facilitate the communities need to take more responsibility in making sure that they’re monitoring for and addressing disinformation. This needs to be supported by community efforts to identify and flag disinformation, to ensure that as few people are uncritically absorbing disinformation as possible.

Unfortunately, a lot of the discourse around disinformation and deep fakes in AI is driven by people who have an agenda to put a halt to AI in its entirety. Or, to be blunt – it’s people concerned about what AI might do to their professional jobs as journalists, lawyers, educators and politicians. 

As a result, a lot of that information being shared is, ironically, misinformation itself. 

There is a clear risk that AI can be leveraged in ways that will be a net negative to society, and spreading disinformation and propaganda can have a meaningful negative impact to everything – from the politicians that we elect, to the information that we get out of conflict zones, when good quality information is critical. 

However, by building better processes to govern how AI is used in the production and oversight of information, we can instead improve the quality of information. 

It just requires humans to understand our ongoing and integral role of proper oversight into the information being generated through these systems.