The world has been on the brink of revolutionary technological change for a while now and Chat GPT and related generative AI technologies might just have tipped us over into a brave new world. Ordinary folks are starting to see the power and possibilities with AI. We are about to see the flourishing of many innovative solutions that are driven by the combined power of data and large language models (LLMs).

But there are some issues facing this technology. For example, I am just waiting for all the copyright, patent, and trademark infringement lawsuits to commence.

My colleague Toby Walsh has been sounding the alarm about the dangers of ‘killer robots‘ – with the danger of autonomous robots being used to attack one’s enemies. For example, Iran’s top nuclear scientist was assassinated by Israeli agents using a robot machine gun in 2020. But let’s just imagine the power of technology that can act autonomously in the real world and talk to generative AI driven business process engines. It will be interesting to see how generative AI will merge with robotics in the real world in the future.

All of this technology is neutral – it can be used for both good and evil. The trouble is that the evildoers can do real damage in the real world. Following is an example of some real world potential impacts arising from generative AI.

New affordances in Generative AI

There was an interesting tweet thread the other day. A user name @NFT_GOD asked “Autogpt to manipulate the results of the 2024 US Presidential election It put together a full scale plan to tear down candidates through wide scale misinformation and deepfake technology The results should scare the hell out of you” – he walks us through his prompts to create an entire misinformation campaign to disrupt the US 2024 Presidential election.

He outlines a process by which one single person can create a disinformation/ misinformation campaign and automate the execution of it. This kind of thing has been done for years, but it used to require armies of humans, and now it can be just bots.

As @NFT_GOD says:

Bots can merge with generative AI

Then there is the problem of bots. Bots can, and probably will, merge with generative AI and this concept holds much promise and peril for human society.

The 2023 Imperva Bad Bot Report, found that automated bot traffic across the internet in 2022 that 47.4% of all internet traffic came from bots. Further Imperva found:

• Bad bots are 30% of automated traffic
• Automated attacks targeting APIs on the rise
• Evasive bad bots accounted for 66.6% of all bad bot traffic
• An in-depth analysis of the threat landscape across industries, sophistication levels, origin and more

So now with generative AI, bots will be able to be created by folks who do not have any special technical skills will then be able to automate entire malware attack chains. We have already seen the allegedly safe Chat GPT used to create malware, where a self-professed novice created a powerful data-mining malware using just ChatGPT prompts, all within a span of a few hours:

“Aaron Mulgrew, a Forcepoint security researcher, recently shared how he created zero-day malware exclusively on OpenAI’s generative chatbot. While OpenAI has protections against anyone attempting to ask ChatGPT to write malicious code, Mulgrew found a loophole by prompting the chatbot to create separate lines of the malicious code, function by function.”

https://www.digitaltrends.com/computing/chatgpt-created-malware/

We are moving into a very interesting phase of world history – it will be fascinating to be a part of this new world (if indeed our society can survive it).