AI Newsletter - June 2024
ProfileTree - Web Design and Digital Marketing
Content Marketing Agency specialised in Website Development, SEO, Content, Video Marketing, Digital Marketing & Training
I’m calling this edition the AI drama edition - because all the big players, Microsoft, Google and ChatGPT have had some sort of PR crisis this month. A humbling moment for the AI industry giants? I think so, but if anything - it also highlights the dire need for human oversight in AI application.
Here’s a quick recap of what we will cover this month.
Buckle up let’s go…
OpenAI is in Hot Water
ChatGPT Founders, OpenAI, is the problem child of artificial intelligence. It seems like every month there is a new controversy surrounding the company and this month they are back in hot water.
Former ChatGPT Board Member Tells All
Ok so, remember when the OpenAI board voted to fire Founder Sam Altam in November last year? Well, one of those board members was Helen Toner. Upon Altman’s dramatic return to the company, Toner was also one of the board members to be let go…
Recently, she was featured on a TedTalk podcast, where she voiced her version of events (or in other words - spilt the tea). In it, Toner revealed that the board was unaware of the ChatGPT launch until they saw it on Twitter, raising questions about transparency and lack of trust in Altman.?
So essentially, we’re getting little snippets of what led to Altman being fired/rehired in November 2023. And look, maybe that’s all it was, but my guess is that it might have been the straw that broke the camel's back for the then-frustrated ChatGPT board.?
Another ChatGPT Announces Resignation then Joins Competitor, Claude.ai?
Jan Leike, former Head of OpenAI Alignment recently announced his resignation on the 15th of May - and just a fortnight later, he has announced his transition to a role at AnthropicAI, the company behind Claude.ai .?
On the social media platform X, Leike highlighted his concerns about OpenAI's commitment to safety, stating, “
“Building smarter than human machines is an inherently dangerous endeavour. Open AI is shouldering an enormous responsibility on behalf of humanity. But over the past years, safety culture and processes have taken a backseat to shiny products”. A somewhat concerning statement to make!
This move is also particularly noteworthy as AnthropicAI was established by former OpenAI employees (and siblings), Daniela and Dario Amodei. Leike’s departure feels like something we would see play out in Game of Thrones, only a modern-day version with robots and computers.
Scarlett Johansson’s Voice Illegally Used by Open AI?
Scarlett Johansson has initiated legal proceedings against OpenAI and its CEO, Sam Altman, alleging unauthorised use of her voice in the latest ChatGPT 4.0 system, named Sky.?
According to Johansson, Altman had initially approached her for permission to use her voice, which she declined. Despite her refusal, Altman made a last-minute plea two days before the release, which she also rejected.
To Johansson's dismay, the voice in the released version of Sky bore a striking resemblance to hers, prompting her to take legal action, which has also led to public backlash. In response to the controversy, OpenAI promptly removed the contentious voice from the system.
Altman has defended the company's actions, stating that the voice was generated by actors and was never intended to mimic Johansson. He explained that the recordings were made prior to their final outreach to her. Nonetheless, out of respect, they decided to eliminate the voice from the system.?
领英推荐
You can view Scarlett’s entire statement below.?
Google AI Recommends Putting Glue into Pizza Sauce
Google recently unveiled a new "AI search" feature, initially available only to U.S. users. However, it was quickly retracted after a significant mishap. What went wrong? Among other questionable recommendations, the AI suggested adding glue to pizza sauce in order to thicken it- a horrible recommendation!
This incident highlights the need for extreme caution when integrating AI into daily applications. While the suggestion to add glue to food is blatantly absurd and easily dismissible, it raises concerns about how AI handles more subtle or commonly known facts.
What Did Google Say?
There have also been instances where users shared images of the AI feature giving hazardous advice, such as endorsing smoking during pregnancy. However, Google has issued a statement clarifying that these responses were fabricated by individuals and not generated by their AI.?
Interestingly, Google's statement did not address the infamous pizza sauce recommendation. This episode serves as a stark reminder of the potential pitfalls of AI, emphasising the importance of rigorous testing and ethical considerations in its deployment.
Microsoft Recall - Gone Too Far?
Microsoft's latest update to its Windows operating system introduces a groundbreaking but controversial feature known as " Microsoft Recall." This new function acts like a comprehensive browser history but for your entire computer.?
It tracks everything you do - whether you're editing a video in Adobe, emailing on Outlook, using Paint, browsing the web, etc., every activity is logged and accessible through the Recall history. It is similar to how one would view their internet browser history, but instead, it’s for your entire device.?
The Appeal of Recall
The idea behind Recall is to enhance productivity by allowing users to effortlessly navigate through their past activities. Imagine being able to trace back your steps to a document you were working on a week ago or revisit a series of actions you took in a software program without manually searching through files and folders. This could significantly streamline workflows and make multi-tasking more manageable.
Privacy and Security Implications
However, with great power comes great responsibility - and considerable risk. The primary concern with Recall is privacy and security. Although all data tracked by Recall remains local on the computer and doesn't sync to the cloud, which ensures that Microsoft doesn't access or collect this data, however, vulnerabilities still exist.?
For instance, if a computer is lost or stolen, anyone who gains access to it could potentially retrieve the entire history of the user's activities, including sensitive information like passwords and bank details. This feature does not blur out any sensitive data, which could be a significant security loophole.
The UK Investigates
The UK has already initiated an investigation into Recall, reflecting the broader implications and concerns regarding user privacy. Users do have the option to opt out of this feature, and they can delete items from their history, which offers some level of control over personal data.
Final Thoughts
Adopting artificial intelligence requires a responsible approach. When AI systems are implemented, their impact is significantly greater compared to launching a typical website. The consequences can be more severe; if not adequately anticipated, they can lead to real harm, spread misinformation, and ultimately erode trust in the technology.?
However, maintaining a healthy scepticism towards AI is advisable and, in my view, beneficial. I hope you found this edition enlightening. We look forward to sharing more insights with you next month. This article is brought to you by ProfileTree Web Design and Digital Marketing.