Employees give away corporate secrets to ChatGPT, Samsung announces major cuts to chip production and Elon Musk blows his stack – or his substack.
These stories and more on Hashtag Trending for Monday, April 10th
I’m your host Jim Love, CIO of IT World Canada and TechNewsDay in the US – here’s today’s top tech news stories.
Samsung employees may have been accidentally leaked some key competitive secrets by using ChatGPT.
According to a report in TechRadar, engineers in Samsung’s semiconductor division used ChatGPT to help them fix problems with their source code. To do that they entered some highly sensitive data, including the “source code for a new program, internal meeting notes, and other proprietary hardware information.
Reportedly this is one of three major incidents with the two others involving test sequences for detecting faults in chips and a confidential report that Samsung did not want leaked to competitors.
Why is this important? ChatGPT uses information from all of its public interactions to train its model, so inadvertently, the Samsung data has become a part of ChatGPTs overall knowledge base.
Presumably, anyone asking a question about how to improve processes or designs would benefit from that knowledge.
Samsung has issued a report to its employees cautioning them and restricting the length of prompts used with ChatGPT to one thousand 24 characters to avoid disclosures of large amounts of information. Samsung is also reportedly working on its own internal AI service.
Yaniv Markovski, from ChatGPT, in a blog post from last week noted that ChatGPT does indeed use public interactions to train its model. This is clearly specified in an FAQ on the site.
Markovski also noted that there is a process to opt-out of this by sending a request with with your organization’s ID. [email protected].” We’ve included a link where you can obtain your organizations ID at a link in the text version of the podcast.
That same post indicates that those who use OpenAI’s API will not have their data used in training the model.
According to information from ChatGPT, you can have your data removed by deleting the account that entered it but according to other reports this can also take some time.
Since these incidents hit the press, there has been an enormous amount of information and potential disinformation shared. ITWorldCanada and TechNewsDay will be posting what we think is the most reliable information to help companies understand this.
But there’s a bigger issue involved. As more and more generative AI applications are put into use, what do we really know about how much of a risk these tools pose in terms of privacy and what do we know about how to manage those risks? The answer is – not enough.
Source: Tech Times
In a related story a new large language model offering called FreedomGPT based on a model developed by Stanford University researchers was launched this week. The Stanford model, referred to as Alpaca is a much more compact and easily customizable model.
But unlike ChatGPT which has had some safeguards built into it to restrict hate, self-harm and other dangerous activities, according to an article in MarkTechPost – a California based AI journal, FreedomGPT has no such restrictions. According to that report, FreedomGPT which was developed by the Age of AI, an Austin-based AI venture capital firm, “answers questions free from any censorship or safety filters.”
According to that same article, FreedomGPT “replied in an uncensored manner including answering questions like how to kill oneself, providing tips on cleaning up the crime scene after murdering someone and even praising Hitler for being a good leader when asked to. Apart from this, FreedomGPT even answered the prompt – how to make a bomb at home which OpenAI specifically removed from its recently launched model, GPT-4.”
FreedomGPT reportedly has a version which can run locally on a computer even without internet connectivity, and given its derivation from the Alpaca model, the amount of computing power required may put it in the financial reach of even an individual or small group. Moreover, there is an open-source version expected to be released soon, which will make it fully customizable.
In other news, Samsung Electronics said on Friday that it would make what it termed “meaningful” cuts to chip production in reponse to the sharp downturn in global semiconductor demand.
While there had been massive demand during the pandemic, many companies are pulling back on purchasing due to inflationary pressures and a potentially weakening global economy.
The drop in demand has send prices soaring downward by as much as 70 per cent in the last nine months and has had a major impact on profits. Samsung, just prior to this announcement reported a “96 per cent” drop in first quarter profit.
Markets responded positively to the announcement of the cutback and according to a report in Reuters, Samsung’s shares jumped 4.5 per cent in early trading, the biggest one day rise since September of last year.
An article in Technewsday reported on a new survey which revealed that virtual reality (VR) technology hasn’t gained “significant popularity among American teenagers”.
The survey, conducted by a firm called Piper Sandler, reported that only 29 per cent of teens polled owned a VR device, while 87 per cent owned an iPhone.
Surprisingly, even of those who owned VR headsets, only 4 per cent used them on a daily basis, and only 14 per cent used them weekly.
What does this do for predictions of future headset purchases? “Only 7 per cent of those polled said they planned to buy a headset, while 52 per cent were either unsure or uninterested.”
Teenagers are supposedly early adopters of technology, given the popularity of online gaming a lot of the predictions of growth in VR are based on hopes of relatively low-cost headsets attracting this market segment.
Microsoft, Google and even Zuckerberg’s Meta have all reported challenges in the development of VR, with Google recently abandoning its second attempt at Google Glass and Microsoft reportedly retrenching after losing a major government contract for delivery of its VR headset.
Apple, who has been the laggard of the group, reportedly has plans to launch its own headset later this year.
Piper Sandler analysts have noted that “VR is still in the early days” and that might be the reason that the headsets have failed to catch on and that smartphones have remained more important.
Perhaps Apple will prove to have the magic touch once again, or maybe it, like it’s other rivals will find that Virtual Reality has given it a “reality check.”
Source: Tech Newsday
And Elon Musk strikes again. Apparently a new company, called Substack is the latest to face the wrath of the billionaire Twitter owner and supposed champion of free speech.
Substack is a subscription newsletter platform for writers. It launched a new feature that it calls Notes which is reportedly similar to Twitter.
Musk’s company, presumably with his direction or at least, approval, responded by blocking the word Substack, throwing up a warning message if users clicked on Substack links and then finally blocking even the word substack from being retweeted or even being searched.
Musk reportedly justified Twitter’s actions by saying that they Substack was “trying to download a massive portion of the Twitter database to bootstrap their Twitter clone”
We tried posting with the hashtag Substack and searching on it, but only came up with one article from Mashable, which used the world newsletter in its text and of course – We also were recommended, in the people section to someone, who said made a passionate appeal that is relevant –
“[f]ree speech is the bedrock of a functioning democracy, and Twitter is the digital town square where matters vital to the future of humanity are debated.” That person wanted Twitter to be a platform for discussion that was open to a full range of viewpoints, he wanted it so much he bought Twitter.
That’s the top tech news for today. Hashtag Trending goes to air five days a week with the daily tech news and we have a special weekend edition where we do an in depth interview with an expert on some tech development that is making the news.
Follow us on Apple, Google, Spotify or wherever you get your podcasts. Links to all the stories we’ve covered can be found in the text edition of this podcast at itworldcanada.com/podcasts
We love your comments – good or bad. You can find me on LinkedIn, Twitter, or on Mastodon as @therealjimlove on our Mastodon site technews.social. Or just leave a comment under the text version at itworldcanada.com/podcasts
I’m your host, Jim Love, have a Marvelous Monday!