On the eve of the 2023 Paris Peace Forum, the Executive Director of Tech Against Terrorism, Adam Hadley, delivered the following remarks at a dinner at the Elysée Palace hosted by the President of France, Emmanuel Macron.
“I want to ask a potentially contrarian question: there's a lot of excitement about large language models and generative AI, but is this real? Is it really going to transform society in the ways many have promised?
Surely we need to focus on fundamentals in our society as much as anything else, and avoid overemphasising novelty? I was speaking to someone earlier about how, last year, the metaverse was presented as something of grave concern and excitement.
I have two points to make.
The first point is about economic disruption, and the potential for generative AI to cause it. The current answer is that it is likely to cause very little disruption because the vast majority of businesses have extremely poor analytics and data capabilities. AI will only support businesses that already have the data infrastructure for better decision making.
For the technology of generative AI to confer any economic benefit, the basics and fundamentals need to be addressed. The question for policymakers should be - how do we focus on those fundamentals? How can government investment encourage businesses to focus on the basics of digital transformation? Whether you're a supermarket, school, or bank, you won't benefit from AI without first having the right foundations in place, such as a proper data infrastructure. You can't benefit from AI without that.
The second point concerns the potential for weaponisation of the internet. At Tech Against Terrorism, our particular concern at present is not that terrorists are using generative AI, although Hamas and Hezbollah are trying to use generative AI image creation and similar tools.
The more pressing concern is that internet shutdowns, which are a grave problem, are not being addressed by policymakers. Terrorists are going "back to the future" and relying on websites, which are very difficult to take down. We at Tech Against Terrorism track around 300 terrorist-operated websites run by al-Qaeda, ISIS, Hamas, Hezbollah, and others. I would add that Hamas is using one particular messaging app with impunity - what's being done about this?
While it's important to discuss the threats and opportunities of AI, it’s critical to focus on the fundamentals. Mr. President, I know you're a great supporter of the Christchurch Call to Action. I'm pleased to say that on Friday we'll announce a partnership to explore the use of generative AI to improve content moderation, which is a great example of how AI could potentially be harnessed to deliver a substantial benefit.
In conclusion, the promises and threats of AI will always be built on foundations; whether in matters of the economy or in content moderation online, those foundations must always be our focus. Thank you."
The discussion was held under Chatham House rules, but permission has been given for Adam’s remarks to be made available.
Adam Hadley is Executive Director of Tech Against Terrorism, a UN-initiated not-for-profit focused on disrupting terrorist use of the internet, and CEO of QuantSpark, which is an analytics and AI consultancy in London.