Generative AI: Risks and Dangers
If you’re reading this article, there’s a good chance you’ve heard of DALL-E2, the Artificial Intelligence (AI) tool that you type in a few words and the tool creates some funky new art, that is often hilarious and also weird. Is it actually art? Perhaps. That’s a hot topic of debate. But it is what’s known as Generative AI. It’s a significant leap forward and will herald now opportunities. But there are downsides, as with any technology. What are the possible unintended consequences of Generative AI?
Unintended Consequences of Generative AI
Search Engine Warfare: A primary obsession of all marketers is SEO (Search Engine Optimisation), it’s a big business in and of itself. A constant battle between the search engines trying to deliver relevant content and marketers gaming them to be at the top of the page. Imagine the battle that could ensue with Generative AI (GAI) locked in mortal combat with search engines. One part of AI working to figure out the search engine’s algorithms, the GAI writing content furiously to get top rankings. Who loses? The average citizen and consumer. We’re already being hacked by social media companies. This takes it to a whole new level.
Job Losses: One of the most common jobs in the gig economy is content writing, for news media sites as well as for companies. Sometimes SEO oriented, but also for corporate blogs and marketing needs. Good GAI could take out many of those jobs. In some cases GAI might also enhance those roles, but more will be lost than survive in cooperative roles. Legal researchers could also lose their jobs.
Cybercrime: Although phishing attack content, such as those emails from a bank or some rich guy in an African country deciding to give you all his millions, are getting better, they can still largely be spotted as humans learn. Imagine a GAI writing such texts and emails. This puts pressure on the human mind to detect and will cause the cybersecurity industry to develop tools to detect GAI.
Doxxing and Personal Attacks: A GAI could also be used to gather information on an individual and create faked images and videos, along with written content to make an innocent person look like they’ve done horrible things. It becomes so complex, so good, that unraveling it may be almost impossible. Automated cyberbullies.
Scam Phone Calls: As GAI helps improve speech synthesis, so might the quality of automated voice spam calls improve. Mixed with auto-dialer software, it could be a nightmare scenario for people. Perhaps it could be used like a DDOS attack on a server, except it’s your phone that gets bombarded with calls.
There are some powerful applications for GAI in business and in society as a whole. A GAI could assist architects in designing better buildings that would be more environmentally friendly. Design cool new products, help lawyers do better research. Video games could become more complex and enjoyable.
While we can’t determine all the risks, dangers and unintended consequences in advance, industry could consider them in a framework. The obvious ones, such as those above, can have safeguards built in today. They may help to reduce the risks and dangers and mitigate for unintended consequences. But as we’ve seen, most technology companies building major AI applications are hesitant to employ ethics that will impact potential revenues and profits. Governments are struggling to catch up to such rapid advancements in AI, let alone other digital technologies.
Generative AI can bring some significant benefits to humanity, but it, like all technologies, are a double-edged sword. As GAI evolves alongside other AI applications such as NLP and ML, we are going to face some deeply existential questions around what it means to be human, how capitalism will have to evolve, what the new welfare state means and how we regulate these technologies while fostering innovation. These are not small issues.