If 2023 has taught us anything, nobody can predict what will happen next in the world of AI, but after a year of placing stories and comments on some of the biggest AI stories this year, here are our best guesses for what 2024 will bring.

Poor integration

This isn’t to say most AI deployments thus far have been world-beating, but there is a public relations disaster brewing with AI. Currently, teens are using the Home Depot chatbot to answer maths questions because of the shoddy integrations underpinning current Gen AI tools, but this smacks of incompetence rather than malice. As the integrations improve and the models become more lifelike eventually someone will be offended or misled in a way that will leave a business liable

Security Risks

Young people historically are great hackers, their natural curiosity and intuitive grasp of software makes them innately talented at breaking software. Currently, there isn’t too much stored in a proprietary model that you could get out of it, however, eventually, there will be chatbots calibrated to individual companies and an enterprising young person will crack one wide open. Depending on the ineptitude of the integration a broken open chatbot could be a window into a company’s intranet with all the interesting information a business would not want in the public eye

Criminal chatbots

Scammers have been held back by their unsteady grasp of the English language. Most phishing emails are clearly illegitimate because corporate comms, despite their deplorable blandness, are hard to replicate. An AI tool can simplify the process of writing and distributing these emails. The tools to do it already exist – Worm GPT is a good example, It is only a matter of time until someone is able to effectively deploy an unprecedentedly sophisticated phishing attack world wide

A government will get it wrong

2024 is an election year for several countries. The potential risks of deepfaked images and voices are constantly in the news. Some politicians may claim some photos are deepfakes rather than evidence of wrongdoing, while governments may attempt to legislate against the technology. It is unlikely to work, but bans or restrictions will be trialled in some countries in an attempt to stem the tide of disinformation. If 2016 was described as the beginning of ‘post truth’ politics then 2024 could be the start of post reality politics. A situation where voters cannot believe the evidence of their eyes and ears.