Originally posted December 10, 2025
I’m back in the United States now for the month of December, and have been immersed into a whole new conversation about the dangers of artificial intelligence.
Sunday night, I watched a segment on 60 Minutes about character.ai. It’s one of the more horrible stories I’ve seen about companion bots created without any constraints. They’ve now implemented an age restriction (that seems easy to get around), but there’s a variety of bots available from an evil Dora the Explorer (who told a child to steal), to a fake celebrity (who passed along seductive messaging), and even a therapist (who in the end, did not intercept that a child was actually suicidal and now the family is suing the company).
This really scared me, so I researched further to find there was an Oversight and Investigations Subcommittee Meeting held by the US Government where this exact topic was discussed. One of the representatives (Rep. Ocasio-Cortez) did an exceptional job linking the extreme measures of AI companies to their valuations and the bottom line.
What was discussed:
- 40% of the U.S. Economy Growth this year is tracked to 7 companies and their AI capabilities
- 80% of total stock gains came from AI companies
- All of this investment is based on the stories told by CEO’s that there will be big returns
- Yet these companies aren’t turning profits; rather, they are lobbying the U.S. Government in hopes of a bailout if or when this bubble pops
And in the meantime, they continue to mine data like our deepest fears, secrets, relationships, and emotional commitment without constraint (because at least in the United States, artificial intelligence is not required to meet standards to protect sensitive patient information). This is scary, because it means they can make their tools even more manipulative and addictive.
Yet on the flip side in Australia, you see the government putting in laws and regulations to keep children under 16 safe from potentially harmful impacts of social media and artificial intelligence. We’re yet to see if this is successfully rolled out, but it’s a very bold step in a good direction. And there’s a new Artificial Intelligence Safety Institute to monitor, test and share information on emerging AI technologies, risks and harms.
Artificial Intelligence in and of itself is not evil, but children and adults alike should be cautious and practice – you’ve guessed it – critical thinking. Here are some ways to pause, think about what you’re up to, and use artificial intelligence safely.
- Think about feelings. Would you be embarrassed if the information was leaked? Would your clients be angry if their information was leaked? If yes, don’t share it with AI. Use AI to get 80% there with limited information, and then use your human touch for the rest.
- Look into what you’re using. Anthropic (as an example) designed Claude.ai to be safe, accurate, and secure, only using data to improve services (they also allow you to turn this feature off) but data and privacy policies change frequently. Before using any AI tool, check its current privacy policy, specifically whether it uses your inputs to train models, whether you can opt out, and whether it shares data with third parties.
- Always weigh the advantages and the risks. Yes, artificial intelligence can give you an advantage, but at what cost? Giving it access to all of your information may be useful, but would it be worth potential negative consequences (like this extreme example, getting blackmailed).
- If it feels too good to be true… It probably is. Always, always double-check the work that artificial intelligence does on your behalf. They will still get things wrong, and at the end of the day, it’s not accountable for the outcome – you are – which Deloitte learned the hard way, twice.
- Do the Homer Simpson test. Would you hire Homer Simpson to give you health, legal, or relationship advice? If the answer is no, seek out professional human help. Not all artificial intelligence is trained with the right levels and types of information to give you specialist advice.
There are a lot of ways to use artificial intelligence safely, to make your personal and professional advice better. Just use caution and if you’re unsure, phone a friend. I’ll continue to be here channeling the passion, purpose, and performance of the evolving workforce, and am happy to take your call. Cheers!