Google's latest version of AI, once Bard but now called Gemini, is yet another indication of how biased the current batch of AI platforms are. I was going to include a bunch of examples but this has received so much coverage that everyone should have seen it by now. Basically, the product offers anything but a white-skinned person in requested pictures. This gave rise to some short-lived pub games. Many found this amusing but after a while it became obvious that Google has shut Gemini down for re-education.
- It gets worse. Google Gemini has little concept of context across stories. There are potential court cases for defamation against the search engine giant for making connections and associations by ignoring context, such as describing someone associated with heinous behaviour if, say, they were just reporting on it. It would appear that Google's processes for AI is flawed and they need to go back to the drawing board. One telling example is Gemini replying that it was not okay to misgender someone to stop a nuclear holocaust.
- Hot on its heels, the Meta AI engine has also started giving inaccurate results which brings into question the whole concept and reliability of an AI engine. An ideal engine would give just the facts with a balanced answer that allows people to make up their own mind on the data. From there, people could add context to get different perspectives. Not being signed up for X Premium, I don't have access to the Grok engine which is touted to be just this kind of balanced product. It will eventually be released to the public and we will soon see.
- Many people will watch videos and commentary from YouTube and other platforms. Like some, I prefer to watch the original source material so I can have my own reactions. I'm the same with a report on a new scientific paper. I want to also read the original document. Some YouTube style videos are instructive, such as how to get past a level in a game or how to build a folding table or repair your washing machine. Some are entertaining, funny, and some others make you mad or feel soothing. The one subject area appearing more often these days though is people commenting on people having a discussion about something, or worse, people commenting on people commenting on such things.
- I'm not even sure how to describe the latter, as adding your opinion to someone else's for a debate or presentation is the start of the old Chinese Whispers game and is unnecessary as most of them simply agree with the content and do not add anything extra. This generates clicks, likes and I assume income for those posting. Is this valuable? Unless it is providing an alternate viewpoint, possibly not, but for some reason there are more and more of these videos popping up. This is also true for reaction videos with a heading similar to "first time listening to the band Dire Straits". Some of us grew up listening to them or seeing them live. Again, these are popular, not sure why. There is also a positive aspect to such recommendations in your YouTube feed. If you missed a good discussion or song, then this will at least alert you to find the original, watch it, and then make up your own mind.
- From March 2 next year, Microsoft will stop supporting the Windows Subsystem for Android (WSA) available under Windows 11. WSA allows users to run Android applications from the Amazon Appstore. There has been no reason given for this 18 months after WSA went live. If you use such apps, you will need to plan for their non-support early next year. In other Microsoft news, do you remember Clippy the helper added to Windows products in the 90s? The equivalent under Windows 11 is Copilot. This will get even more intrusive as Microsoft has just added a function that causes it to pop up when the mouse cursor hovers over an icon in the task bar. Hopefully you can turn it off in Settings, Personalization, Taskbar. As a side note, Windows 10 is still well ahead of Windows 11 in terms of adoption and usage stats. It is less than a year-and-a-half until the official Windows 10 retirement so Redmond has a lot of work to do to get people to shift across.
- On the subject of Amazon, they have just agreed to forego exit fees for those with lots of data leaving the AWS platform. They join Google in this after Europe enacted regulatory requirements covering this kind of thing.
- Supporters from both sides of politics are creating AI generated deepfakes to boost their candidates. As predicted, 2024 is the year deepfakes that appear real, can get a million views and can paint a potentially false view of reality. With both sides creating them, it will be interesting to see if any steps are taken to stop this.
James Hein is an IT professional with over 30 years' standing. You can contact him at jclhein@gmail.com.
