I'm a yuge fan of AI - just not the politically crippled, malevolence-enhanced versions that have been publicly released.

"One of the most promising applications of AI is in the field of healthcare."

Indeed, a properly configured AI app can mostly replace doctors' diagnoses. A lot of diagnoses, like most identification of species, is little more than a checklist. If you've ever used an Audubon or other field manual, you'll know exactly how that works. Because of cladistics, it's a relatively simple matter to differentiate one species from another, and for pathogens as well as daisies, checklists of features enable that identification to be undertaken. Certainly there's more to diagnosis of medical conditions, and doctors themselves get things wrong, which is one of the reasons that iatrogenic causes were the third leading cause of death in America prior to the jabs, and are now the leading cause of death, because of the jabs.

One of the most exciting uses for AI IMHO is automation of decentralized production. As computer hardware increases in capacity while decreasing in cost, AI scales ever smaller far better than it scales larger, just as means of production are today. The learning curves for many if not all of individually applicable means of production are likely to be eliminated pretty soon. Managing aquaponics, household robotics, 3D printing, power management, and most systems that the IoT now forces people to upload their data will soon be handled in house, ending that data harvesting and surveillance capitalism.

The more decentralized AI gets, the less hazard it presents. It is only now, while it is the exclusive property of transnational corporations, that malicious AI presents much hazard. Since LLM's are already able to be downloaded, and people I know are in possession of GPUs on which small AI can run, that corporate installed malice will quickly be eliminated as a problem we need to face.

A couple years ago facial recognition software similarly escaped corporate corrals, and a program that identified the social media accounts of porn actresses was released, causing quite a furor. However, after the Jan. 6 hearings and vilification of protesters, facial recognition software was applied to the videos by independent people to identify state actors and undercover cops inciting violence and crimes. It won't take long before covert actors and undercover cops are a thing of the past, because facial recognition is no longer exclusively a power the state and transnational corporations have, but today is wielded by private individuals.

However, our complete inability to even define consciousness, know where it comes from, how it arises, or why, makes the specter of rogue AI impossible. AI is, in fact, a misnomer. What we actually have seen and have any capability of making are complex algorithms, and that has nothing whatsoever to do with intelligence.

"...chatbots and other AI-powered tools can improve customer service..."

From what I have seen to date, AI chatbots deployed in customer service roles significantly degrade service, and probably increase expenses, because in addition to the cost of the AI, actual humans that can resolve issues still need to be maintained. Every single time I try to use such a service I am compelled to resort to human beings to resolve issues, some of which are caused by chatbots themselves.

Thanks!